International Journal of Production Research, 2020

Vol. 58, No. 7, 2222–2241, https://doi.org/10.1080/00207543.2019.1683248

A blockchain-based approach for a multi-echelon sustainable supply chain

V. K. Manupatia , Tobias Schoenherrb , M. Ramkumarc , Stephan M. Wagner c∗
, Sai Krishna Pabbad and
R. Inder Raj Singhd

a Department
of Mechanical Engineering, National Institute of Technology Waraal, Warangal, India; b Department of Supply Chain
Management, Broad College of Business, Michigan State University, East Lansing, MI, USA; c Chair of Logistics Management,
Department of Management, Technology, and Economics, Swiss Federal Institute of Technology Zurich, Zurich, Switzerland;
d Department of Mechanical Engineering, Vellore Institute of Technology, Vellore, India

(Received 21 December 2018; accepted 22 August 2019)

Blockchain technology is destined to revolutionise supply chain processes. At the same time, governmental and regulatory
policies are forcing firms to adjust their supply chains in response to environmental concerns. The objective of this study is
therefore to develop a distributed ledger-based blockchain approach for monitoring supply chain performance and optimising
both emission levels and operational costs in a synchronised fashion, producing a better outcome for the supply chain. We
propose the blockchain approach for different production allocation problems within a multi-echelon supply chain (MESC)
under a carbon taxation policy. As such, we couple recent advances in digitalisation of operations with increasingly stringent
regulatory environmental policies. Specifically, with lead time considerations under emission rate constraints (imposed by
a carbon taxation policy), we simultaneously consider the production, distribution and inventory control decisions in a
production allocation-based MESC problem. The problem is then formulated as a Mixed Integer Non-Linear Programming
(MINLP) model. We show that the distributed ledger-based blockchain approach minimises both total cost and carbon
emissions. We then validate the feasibility of the proposed approach by comparing the results with a non-dominated sorting
genetic algorithm (NSGA-II). The findings provide support for policymakers and supply chain executives alike.
Keywords: Blockchain; production allocation problem; multi-echelon supply chain; smart contracts; carbon taxation
policy; Mixed integer programming

1. Introduction
Concerns for the environment have become front and centre for companies worldwide (Kumar, Subramanian, and Ramku-
mar 2018; Ramkumar and Jenamani 2015). While there are many drivers for this increased environmental consciousness,
one critical determinant has been the rise in environmental regulations in general, and the ratification of carbon taxation
policies specifically (the focus of this research). Hence, mitigating and reducing carbon emissions has become an impor-
tant issue for companies, enabling them to be environmentally responsible while reducing the economic impact of carbon
taxation (Manupati et al. 2019).
However, it is not enough for companies to focus on their own internal operations; they must also be mindful of their
wider supply chains, since actors responsible for the focal company’s input materials can also have a significant impact
on the company’s overall performance. Supply chain entities thus not only need to be concerned about themselves, but
also about their supply chain partners. This applies to all aspects of operations, including the environmental dimension,
and emphasises the need to look at the problem in an integrated and multiplex fashion, considering the complexity of the
environment. Thus, companies have to move away from the myopic view of merely managing their own environmental
impact, and towards the management of their suppliers’ environmental impact (Sundarakani et al. 2010; Tognetti, Grosse-
Ruyken, and Wagner 2015). Environmental concerns should thus be embedded within companies’ supply chains to mitigate
environmental risks (Bové and Swartz 2016). This becomes evident, for instance, in the increasing popularity of holistic
environmental supply chain certifications, with the chain of custody certification by the Forest Stewardship Council (FSC)
serving as one example (Narasimhan et al. 2015). Certifications like these are evidence of the sustainable management of
input materials throughout the supply chain.
Within this context, we provide a framework for assessing carbon emissions across the supply chain, rather than within a
specific company. We focus on carbon emissions, since many countries have committed to the reduction of carbon emissions

*Corresponding author. Email: stwagner@ethz.ch

© 2019 Informa UK Limited, trading as Taylor & Francis Group

 International Journal of Production Research 2223

by the year 2020 by applying a number of protocols (Liu et al. 2017; Manupati et al. 2019; Yang, Wang, and Shi 2017). The
importance of developing additional insight is given by the fact that current information on carbon emissions is scant and
thus potentially misleading (Chaabane, Ramudhin, and Paquet 2012; Manupati et al. 2019). As such, in addition to enabling
companies in undertaking proactive decisions and taking a responsible stance on sustainability, the devised approach focuses
on attaining profitable margins whilst considering environmental aspects. The optimisation of carbon emissions alongside
the associated operational costs distinguishes this methodological approach by considerably reducing total expenses across
the supply chain.
To address the complexity created by the multitude of players in a supply chain and the information that they can provide
for an optimal management of a company’s environmental performance, we leverage technological advances inherent to a
distributed ledger-based blockchain approach in a consortium blockchain, minimising both total cost and carbon emission
levels, and optimising overall supply chain performance. We position the framework as a mechanism that can monitor
various functions of the supply chain in an integrated manner, and thus enable real-time decision making in a collaborative,
multi-echelon environment. In this kind of environment, technology seems promising, as the reduction of emissions in
each stage of the supply chain will reduce the total emissions for a focal company (Fu, Shu, and Liu 2018). Due to its
tracking abilities for all stakeholders, blockchain technology can also make complex, multi-echelon supply chains more
transparent and efficient (Galvez, Mejuto, and Simal-Gandara 2018; Saberi et al. 2019; Schmidt and Wagner 2019). Previous
transactions, entities involved, time spent at any one node, and input sources for the final product can, for example, be traced
and verified. Root cause analysis can thus be performed in minutes. In addition, a blockchain’s probabilistic approach with
thousands of miners make it a trustworthy and secure platform. As a result, blockchain technology can simplify complex
systems by enabling visibility and control, thereby reducing cost and time and increasing the efficiency of the entire system
(Galvez, Mejuto, and Simal-Gandara 2018).
We devise a multi-objective decision-making framework for designing an efficient supply chain, a requirement of all
modern-day operations (Banasik et al. 2018). Our framework is unique because of the integration of lead time constraints,
which prior work emphasised as critical (e.g. Banasik et al. 2018; Hammami and Frein 2014; Hammami, Nouira, and Frein
2015). In addition, research has offered only simplified versions of reality, by, for instance, not considering raw materials
and intermediate products when computing final product emission values (Manupati et al. 2012). Further, operations man-
agement scholars have not extensively considered the carbon footprint of supply chains (Benjaafar, Li, and Daskin 2013).
Our work addresses these shortcomings.
In addition, present studies in the field of sustainable supply chain management reveal that, in spite of redesigning,
optimising and simulating supply chains, attaining ideal solutions remains a major challenge (e.g. Hammami and Frein
2014). Very few models optimise the variables under consideration in this research – carbon emissions and operational costs
– by making either of them a primary concern. This study explores the optimisation of both variables using a self-developed
framework capable of yielding maximum profits. The ‘consortium blockchain’ framework is implemented in the multi-
echelon supply chain model, which enables live monitoring of the supply chain. The live data inspection helps in decision
making by allocating suitable production management resources for achieving system optimality.
Furthermore, recognising that companies cannot focus solely on environmental performance, but are certainly also con-
cerned with overall performance, including cost and efficiency, dimensions need to be weighted using suitable constraints;
such constraints are based on assumptions that need to be developed, ultimately affecting the parameters. However even
then, decisions pertaining to production and inventory, while apparently unrelated to the environment, can have an indirect
effect on the company’s environmental performance (Benjaafar, Li, and Daskin 2013; Bouchery et al. 2012; Hua, Cheng,
and Wang 2011; Wahab, Mamun, and Ongkunaruk 2011). Our research considers this multi-dimensional setting.
This study systematizes the production assignment problems for a sustainable multi-tier supply chain, and recommends
the integration of a carbon taxation policy into the supply chain network with an automated monitoring and data handling
mechanism. The imposition of a carbon tax has been a critical element for sustainable supply chain management (Choi
2013). In addition, we evaluate the competing objectives of total cost optimisation and carbon emissions reductions among
supply chain members, leveraging recent advances in blockchain technology. We then validate the viability of the proposed
method by equating the results with a non-dominated sorting genetic algorithm (NSGA-II). In this way, we contribute to
the literature on the production distribution problem by devising a mechanism that can monitor the entire supply chain and
assist in optimal decision making. The problem is enhanced by the ability to consider the impact of both lead time and
safety stock constraints, as well as differences in regular and overtime production rates. The problem is designed to address
the trade-off between net operational cost and carbon emissions. As a whole, this research illustrates decisions regarding
production, distribution and inventory control in conjunction with carbon emission considerations for a multi-tier supply
chain, utilising a distributed ledger-based blockchain approach.
The article proceeds as follows. After a literature review covering production and inventory optimisation in a supply
chain context under carbon taxation policy, a detailed description of the problem along with its assumptions is provided.
2224 V. K. Manupati et al.

This is followed by the development of a mathematical model and its constraints. The framework of the proposed distributed
ledger-based blockchain is then presented. The section that follows explains the blockchain process with respect to the
optimisation of a multi-tier sustainable production distribution supply chain system, specifically considering lead times
under carbon taxation policies. The ensuing section describes the experimentation conducted with the model, followed
by an illustration of how the decision variables and objective functions are constructed in accordance with influencing
parameters, assumptions, and constraints. The results and their discussions are then explained. Conclusions are provided in
the final section.

2. Literature review
2.1. Carbon taxation and supply chain management
Many countries have made dedicated plans to reduce their carbon emissions by the year 2020 through the application of
various protocols (Liu et al. 2017; Manupati et al. 2019; Yang, Wang, and Shi 2017). It has therefore become imperative for
many organisations to redesign their supply chains to reduce carbon emissions at each stage of the supply chain. Reducing
emissions at each stage of the supply chain is crucial in reducing emissions overall (Elhedhli and Merrick 2012; Jaber et al.
2013; Mathiyazhagan and Haq 2013). However, the literature has largely neglected the cause of carbon emissions across
the supply chain, particularly as it relates to the production, distribution and allocation problems of supply chain design
(Benjaafar, Li, and Daskin 2013; Hammami, Nouira, and Frein 2015); only recently have scholars started to investigate this
domain (e.g. Chaabane, Ramudhin, and Paquet 2012; Fahimnia et al. 2015; Zakeri et al. 2015). Most of the literature only
considered a simplistic supply chain structure focusing on the emissions generated from the finished products, rather than
taking into account the emissions arising from raw materials and intermediate products. The need to address these issues
related to multi-echelon sustainable production–distribution supply chains with lead time considerations under a carbon
taxation policy is thus urgent (Fard and Hajaghaei-Keshteli 2018).
This neglect of carbon emissions across the supply chain is surprising, since sustainable supply chain network design and
planning has garnered significant attention from economists, environmentalists, consumers, industrialists, governments and
academia, because of its strong link to organisational competitive advantage (Letmathe and Balakrishnan 2005; Manupati
et al. 2019; Tognetti, Grosse-Ruyken, and Wagner 2015; Varsei and Polyakovskiy 2017). As such, organisations are delving
into more complex supply chain design and planning decisions (e.g. logistic network configuration and re-configuration,
mergers and acquisitions, outsourcing, reshoring and offshoring) to gain competitive advantage, both economically and
environmentally (Babazadeh et al. 2017; Manupati et al. 2019; Samadi et al. 2018). The importance of considering multi-
echelon supply chains within the context of carbon emissions was emphasised by Hammami, Nouira, and Frein (2015),
who offered insight into the effect of individual emissions caps on each facility when compared to a global cap on the
entire supply chain. One explanation for the lack of research in this domain is the complexity of such optimisation models,
especially when other aspects are taken into account, such as lead time constraints (Hammami and Frein 2014). Lead time
considerations in multi-echelon supply chains are important to account for, since they can have significant impact on carbon
emission inventories (Benjaafar, Li, and Daskin 2013).
Due to rapid industrial development and recent elevations in carbon emissions elevations, there has been a need to assess
carbon taxation policy in the supply chain across the three pillars of the triple bottom line (economic, environmental and
social) (Babazadeh et al. 2017; Samadi et al. 2018). The incorporation of sustainability as part of the triple bottom line
into supply network design often led to variations in the optimum amount of manufacturing activity and storage alloca-
tions, decisions associated to the location of each resource, and the capacity of individual resources (Eskandarpour et al.
2015; Manupati et al. 2019). Deterministic models, capable of incorporating carbon emissions into inventory models under
multiple varying parameters, offer a formidable avenue to investigate such complexities with the ability to optimise carbon
emissions and operational costs simultaneously (Hammami, Nouira, and Frein 2015; Manupati et al. 2019). There is a need
for more insight into how to best deal with this challenging context.
Recent work at the intersection of carbon taxation and supply chain management includes Fahimnia et al. (2015), who
developed a model for preemptive supply chain planning that integrates the dual goals of economic and carbon emission
under carbon tax policy and solves it with a Cross-Entropy method. Insights generated include the impact on financial
and emissions results, an illustration of the use of cost/emission tradeoff analysis, and guidance to price carbon for best
environmental returns. Chaabane, Ramudhin, and Paquet (2012) devised a multi-period Mixed Integer Linear Programming
framework for supply chain lifecycle analysis, considering the trade-offs between economic and environmental objectives,
and calling for a strengthening of legislation. Tseng and Hung (2014) proposed a strategic decision-making model capturing
both operational and social cost dimensions due to carbon dioxide emissions within the context of an apparel manufacturing
supply chain network. We agree with Chaabane, Ramudhin, and Paquet (2012) that there is a need for better legislation. Choi
(2013) considered a supplier selection problem in the presence of different carbon emission tax formats, and Zakeri et al.
 International Journal of Production Research 2225

(2015) developed an analytical supply chain planning model that considers carbon pricing (taxes) and carbon emissions
trading. Zakeri et al. (2015) find that a carbon trading mechanism results in better supply chain performance than a carbon
tax, although the latter may also be advantageous at times.

2.2. Blockchain technology and sustainable supply chain management

Supply chains are becoming increasingly complex due to the interplay of many more actors than in the past (multi-echelon),
and geographically disjointed entities competing to serve customers (Bode and Wagner 2015; Saberi et al. 2019). This,
however, increases the risk that organisations face, since so much more can go wrong in longer and more dispersed supply
chains, because of, for instance, varied government policies, and diverse cultural and human behaviours in these supply
networks (Fu, Shu, and Liu 2018; Ivanov, Dolgui, and Sokolov 2019; Saberi et al. 2019). This often leads to a need for better
information sharing, visibility and verifiability in the supply chain; complications can, for example, result from ineffective
and cumbersome dealings, deception, pilferage and poor coordination among supply chain entities. Blockchain technology,
a distributed ledger-based technology that enables transparency, traceability and security in supply chains, promises to be
a valuable capability for addressing problems in the global supply chain (Queiroz and Wamba 2019; Saberi et al. 2019;
Schmidt and Wagner 2019).
Blockchain technology can be defined as a ‘distributed database of records, or public ledger of all transactions or digital
events that have been executed and shared among participating parties’ (Crosby et al. 2016, 7). Blockchain technology con-
sists of a series of cryptographic time-stamped immutable records of blocks that are consecutively interlinked. A timestamp
includes identification details to indicate an occurrence of a certain event; in our case a transaction. The ledgers prepared
for each transaction with encrypted digital signatures of the parties make it secure. Copies of data are distributed amongst
the participants (including miners) of the blockchain transaction and are interlinked, making the data immutable. Once the
miners have validated the transaction, it becomes part of a distributed ledger, with the miners receiving a credit of cryp-
tocurrency in return for their validation. The blocks possess digital signatures that are unique and carry the ledgers within
by means of a cryptographic time stamp. Moreover, each block contains a cryptographic hash of the previous block, mak-
ing it impossible for an external entity to breach the system; this allows for the distribution of digital information without
alteration. The connected blocks make it easy to trace a transaction through the ledger’s unique digital signature.
Blockchain technology has the potential to disrupt the supply chain because of its trustworthiness, the legitimacy of
information, and its capability to facilitate smart contractual relationships (Saberi et al. 2019). Although blockchain use has
increased in the last few years, the technology faces several behavioural, organisational, technological or policy-oriented
barriers to its implementation in supply chain networks (Crosby et al. 2016; Queiroz and Wamba 2019; Saberi et al.
2019; Schmidt and Wagner 2019). The incorporation of blockchain technology into the supply chain improves traceability,
auditability and verifiability of each stage in the production.
The ‘consortium blockchain’ is the most applicable framework for a multi-echelon supply chain considered in this
research. A consortium blockchain is semi-private and is controlled by different participants in a supply chain in a user group
on a common platform. The infrastructure enables a better workflow, transparency, traceability, visibility and predictability,
for instance to develop robust forecasts. In addition, the environment is controlled as the shared data remains within the
frame of the consortium. At the same time, all authorised stakeholders are interconnected and data sharing is facilitated
(Galvez, Mejuto, and Simal-Gandara 2018).
Blockchain technology can be implemented in supply chains by smart contracts that facilitate, verify and enforce the
agreement (transaction) between supply chain participants automatically. This provides a ‘proof-of-work’ and allows true
data sharing in terms of distributed ledgers. The implementation of a tailored cryptocurrency can also be used to replace the
currency transaction with a data transaction. This data can be used for traceability in a transparent and trustworthy manner.
Specifically, in our context, carbon emissions data can be traced via transactions of carbon assets (digital currency) under a
pre-programmed smart contract that triggers the system upon reaching a carbon emissions threshold limit at the nodes. A
transaction beyond the specified carbon credit threshold limit will therefore trigger the pre-programmed smart contract to
initiate actions to optimise the system. These transactions present supply chain managers a record of data through which the
entire system can be monitored and altered accordingly, as a result of which the carbon emissions can be minimised and the
system optimised.
With these properties, blockchain technology seems most promising for sustainability, which has been attributed to be
a central reason behind the ‘hype’ around blockchain technology (Cole, Aitken, and Stevenson 2019). The intersection of
blockchain technology and sustainability was also illustrated in a recent review by Garcia-Torres et al. (2019). Another
study reviewed blockchain technology and smart contracts, in addition to their application to sustainable supply chain
management (Saberi et al. 2019). With this technology, according to the authors, several barriers to more sustainable supply
chain management can be overcome. Tracking and tracing capabilities, enabled via blockchain technology, can be helpful
2226 V. K. Manupati et al.

(Tian 2016), and have received numerous applications. For example, blockchain technology was seen as beneficial within the
context of sustainable water management (Zhao et al. 2019), the sustainable agriculture supply chain (Kamble, Gunasekaran,
and Gawankar 2020; Sharma, Kamble, and Gunasekaran 2018), and the decarbonisation of product supply chains (Koh et al.
2013).
The importance of traceability for sustainability and the potential of blockchain technology has been elevated by firms’
objective to be perceived favourably by the public, to satisfy the requirements of environmental activists, and to retain their
customers. These objectives have, in some instances, even become more prevalent than an emphasis on cost minimisation
or profit maximisation (Arampantzi and Minis 2017; Devika, Jafarian, and Nourbakhsh 2014; Eskandarpour et al. 2015;
Kannegiesser, Günther, and Autenrieb 2015), or have at least been considered as a trade-off between sustainability and cost
(Tognetti, Grosse-Ruyken, and Wagner 2015). In light of these realities, our research addresses an area of great concern:
how to ensure a firm’s sustainability while remaining concerned with the cost imperative.

3. Problem description and mathematical model

In response to the dynamic and frequently changing environment to which today’s supply chains are exposed, we construct
an adaptive algorithm to determine the workloads of vendors with respect to their capabilities and calculate their total
optimal cost. In our case study, we consider a three-echelon production allocation model, also considering environmental
emissions. We consider two suppliers, three manufactures and four distributors. Items I are moved from supplier S to
manufacturing unit M, with M then forwarding the processed goods P to the distribution centre W, as shown in Figure 1.
Consumers and corresponding distribution centres have stochastic demands. Several time periods are considered in the
planning horizon. This represents a sequential three-echelon supply chain system, where raw materials and products flow
from suppliers to distributors.
The proposed model is solved by leveraging a blockchain approach to optimise both the operational demands and the
carbon emissions; the quantities of carbon emissions at each outlet act as the transaction medium. The amount of carbon
emissions (carbon credits) are denoted by carbon assets in our system and are continuously transferred from the sender’s
account to the receiver’s account in line with the goods transfer. Smart contracts are used for each individual party to monitor
every transaction and store the data of carbon emissions in the form of a decentralised distributed ledger of its outlet. Each
outlet has a limit; if any sudden change is observed or if the threshold limit is surpassed, the smart contract triggers an alert
stating that the emission in the corresponding outlet is unstable.
The updated data from the respective outlet is taken as an input for the optimisation of the system using our proposed
algorithm described below. Subsequently, changes in operational demands must be made to stabilise the system. At this
point, the smart contracts will start the analysis by examining the emissions in the respective outlet and the iterations
continue.
Next, we develop a Mixed Integer Non-Linear Programming (MINLP) model for calculating the total optimal cost with
emissions, considering inventory and safety stocks, and leveraging blockchain technology. The objective is to minimise

Figure 1. Multi-echelon supply chain.

 International Journal of Production Research 2227

the total cost and emissions at each outlet. The objective function is built by defining assumptions, parameters and decision
variables of the model. The objective function with these constraints is then solved with the following assumptions (notations
are summarised in Table A1 in the Appendix):

(1) The lead time of item I at plant M is a constant parameter.

(2) The parameters in the entire supply chain network have a standard normal distribution.
(3) There are no constraints on order quantity, and capacity limitations are neglected.
(4) Demands of participants in the supply chain are fulfilled.
(5) Assembly times include setup times. Since the demands are fulfilled, there is no possibility of an item shortage
while the final product is being assembled.
(6) An output product, which may either be an intermediate or end product, is manufactured by facility M (M ∈ M1,
M2, M3). Each section of this output product may be obtained from multiple sections of distinct input products.
(7) Facility M may have several nodes towards the source, which might include other manufacturing sectors or external
suppliers. It is assumed that each facility M has one node towards the source for each input product.
(8) The emission rate is not fixed, and is monitored for each outlet.
(9) The transportation emission between any two outlets is considered as the emissions of the initial location (i.e. if
the goods are transferred from supplier S to manufacturer M, the transportation emission encountered during the
process is the emission of the supplier S).
(10) For ease of interpretation, we assume that the smart contracts are triggered in the case study (i.e. S = 1).

The proposed algorithm is presented in Equation (1) and consists of several parts. The first part of the objective function
indicates the fixed total costs with fixed emissions. The second part refers to the ordering and total holding cost at each
manufacturing site based on the economic order quantity (EOQ). The third part specifies the calculations based on the
reorder point, including any buffer or safety stock. The fourth part accounts for the inventory holding cost along with the
emission cost (where emission cost is the product of emission quantity and carbon taxation based on the policy). The fifth
part refers to the total manufacturing cost and the total emission cost due to manufacturing. The sixth part signifies the
transportation cost from S to M and its associated emission cost. The seventh part represents the transportation cost from M
to W and its associated emission cost.
Objective function:
Minimise Z, where

⎧    ⎫
⎪
⎪ [(FM ∗ XM ) + [(EFM ∗ XM ) ∗ τ ]] + 2(HMI + EIPMt ∗ τ ) ∗ OMI ∗ DMI ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎪ M M I ⎪
⎪
⎪
⎪     ⎪
⎪
⎪
⎪ + (HMI + EIPMt ∗ τ ) ∗ Z1−α ∗ LMI ∗ VMI + μWM ∗ (HWP + EIPW ∗ τ ) ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎪
M I C P t ⎪
⎪
⎪
⎨  n ⎪
⎬
Z= S∗ + [QRWMPt ∗ CMPt + QOWPMt ∗ CWPMt ] + [EVM (QRWPMt + QOWPMt ) ∗ τ ] (1)
⎪
⎪ ⎪
⎪
⎪
⎪ 
W M P t M =1 ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎪ + [TMSI ∗ DMI ∗ ZMSI + (TEFIS ZMSI + TEVIS DMi ) ∗ τ ] ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎪
M

S I
⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎩ + [TWMP ∗ μWM ∗ YWMP + (TEFPM YWMP + TEVPM μWM )) ∗ τ ] ⎪
⎭
W M P

The purpose of this function is to reduce the total optimal cost and carbon emissions of the entire system. The model
accounts for costs of inventory, transportation, ordering and manufacturing. Both regular and overtime manufacturing costs
are considered, as is safety stock and the respective emissions.
The inventory is stockpiled in the distribution centre to satisfy customer demand with an expected probability of (1-α)
during the lead time period LMI . Given inventory service levels, this probability function is determined as presented in
Equation (2).

Pr(D(LMI ) ≤ rMI ) = 1 - α (2)

where D is the demand of item I during the lead time L in outlet M.

2228 V. K. Manupati et al.

The reorder point is calculated considering the safety stock and assuming a normal distribution. This is reflected in
Equation (3).

rMI = E(DMI ) ∗ E(LMI ) + Z1−α E(LMI ∗ VMI ) + E(DMI )2 σLT

2 (3)

where σ LT is the variance for the demand in the lead time period. While considering a constant lead time, we can neglect
the variance and the reorder point, and note the following:

rMI = DMI ∗ LMI + Z1−α VMI ∗ LMI (4)

Here, Z1−α is indicated as the standard normal distribution value, which is uniform across the network. Based on
Equation (4), the inventory holding cost is calculated by Equation (5), which includes information on the average holding
cost, the EOQ and the cost incurred by holding safety stock.
 
HMI ∗ QMI /2 + HMI ∗ Z1−α ∗ LMI ∗ VMI (5)

Thus, the total holding and ordering costs are represented as follows:
   
OMI ∗ DMI /QMI + HMI ∗ QMI /2 + HMI ∗ Z1−α ∗ LMI ∗ VMI (6)
M I M I

As per the third assumption noted above, capacity limitations are not considered. Setting Equation (6) to zero yields an
equation in terms of Q, which results in Equation (7).

(2 ∗ OMI ∗ DMI )
QMI = (7)
(HMI )

On combining Equations (6) and (7), the aggregate production allocation model is formulated to decrease the total optimal
cost of the supply chain system. Here, Equation (8) ensures that the facility of supplier S is open such that it provides items
I to satisfy the demands of manufacturing unit M.

ZMSI = XM ∀I = 1, . . . , I ∀M = 1, . . . , M (8)
S

Equation (9) ensures that all distribution center (warehouse) demands are fulfilled for all products by a single operating
plant, and YWMP states that products P are transported from manufacturing plants M to warehouses W.

YWMP = 1 ∀W = 1, . . . , W ∀P = 1, . . . , P (9)
M

Constraints (10) and (11) denote the capacity shortage and production limitations, respectively, encountered by manufactur-
ing plant M.
Here, ScapM represents the supply limitation for manufacturing plant M, and PcapMP represents the production limitation
for product P from manufacturing plant M.

DMI ∗ SI ∗ ZMSI ≤ ScapM ∗XM ∀M = 1, . . . , M (10)
S I


μWP ∗ TP ∗ YWMP ≤ PcapMP ∀M = 1, . . . , M (11)
W P

Equations (12) and (13) are applied to generate the mean and variance, respectively, for product P, which is to be
manufactured at plant M.

μWP ∗ YWMP ∗ bPI ≤ DMI ∀M = 1, . . . , M ∀P = 1, . . . , P (12)
W I


σWP ∗ YVMP ∗ b2PI = VMI ∀M = 1, . . . , M ∀P = 1, . . . , P (13)
W I
 International Journal of Production Research 2229

In Equation (14) X, Y and Z are binary variables, taking on values of either 0 or 1.

XM , YWMP , ZMSI ∈ {0, 1} ∀I = 1, . . . , I ∀W = 1, . . . , W ∀P = 1, . . . , P

∀M = 1, . . . , M ∀S = 1, . . . , S (14)

Constraint (15) is used to balance the demands for each distribution centre by considering inventories of the current and the
previous period. It also balances the quantity to be produced for each product P.

LnWP(t−1) + QRWMPt = LnPt ∀W , M , P (15)

Equations (16) and (17) state the production quantity constraints for regular and overtime manufacturing hours,
respectively.

QRWMPt ∗ TP ≤ TRMt ∀M (16)
W P


QOWMPt ∗ TP ≤ TOMt ∀M (17)
W P

Equation (18) reflects the storage capacity of the warehouse.


LnWPt ∗ UP ≤ ScapW ∀W , t (18)
W

Constraint (19) ensures that product P is produced at opened plant M.


[QRWMPt + QOWMPt ] ≤ YWMPt ∀W , M , P (19)
t

Constraint (20) verifies that both regular time and overtime production are always greater than zero.

QRWMPt , QOWMPt ≥ 0∀W , M , P, t (20)

The functioning of facilities related to supplier S is validated by Constraint (8). All distribution centre demands for all
final products are satisfied by a single opened plant, a Constraint that is ensured by Equation (9). The authentication of the
production restriction and storage capacity for plant M is implemented in Constraint (10) and (11). The validation of the
variance and the average number of the products produced at plant M is done by Equations (12) and (13), respectively.
To make these units function continuously, the status of the binary variables X, Y, Z is always considered to be 1, as
per Constraint (14). The balancing of the distribution centre demands in conjunction with the inventories of the previous
period, the current period and its production quantity for each product P, is ensured by Constraint (15). The limitation of
production quantities during regular and overtime hours is reflected in Equations (16) and (17), respectively. Equation (18)
indicates the storage capacity of the warehouse, and the effective production of product P at opened plant M is ensured by
Constraint (19).

4. A blockchain framework
4.1. Overview
We aim to leverage the power of blockchain technology and apply it to the model derived above, seeking to decrease total
cost while minimising carbon emissions. Figure 2 depicts the proposed blockchain framework for a multi-echelon produc-
tion distribution supply chain system that can optimise the relationship and potential trade-off between total operational cost
and carbon emissions. In this framework, the multi-echelon supply chain consists of suppliers, manufacturers, and distribu-
tors under a central manager. Given that a goods transfer and the associated carbon emissions occur between a manufacturer
and a distributor, the direction of movement for both assets is from the manufacturer to the distributor. This ensures the
creation of a track record for all emissions data in the form of a distributed ledger.
The framework explains the mechanism of the system and illustrates how the blockchain is linked with the supply chain
through the help of smart contracts. In case there is a malfunction in the supply chain at a particular node, the smart contract
triggers the user to balance the current scenario for optimisation using the data obtained from the blockchain. The function
of the algorithm is cost minimisation, wherein the carbon emission quantity is a variable and must be converted to monetary
2230 V. K. Manupati et al.

Figure 2. Blockchain framework.

value by multiplying emission quantity with carbon tax to obtain emission cost. The developed algorithm functions after
the conversion of carbon emission units to cost for the supply chain optimisation. Such a functionality cannot be achieved
without inclusion of the blockchain. The optimising mathematical algorithm minimises the cost and achieves results better
than before. All these features are supported by blockchain technology.
In sum, Figure 2 outlines how goods are being transferred, and how carbon assets are transferred between two outlets.
This illustrates the role of smart contracts in the system and the working mechanism of the blockchain. The manager is the
head of the supply chain system, initiates system changes based on conditions set in the smart contract. We now discuss the
procedure in detail, along with the associated algorithm.

4.2. Blockchain mechanism

Considering a scenario in which a transaction is made between a manufacturer and a distributor, the manufacturer, while
performing a transaction on a blockchain, generates two keys, one private and the other public. The public key is used
by the public layer (miners) and other stakeholders of the supply chain to monitor the transactions, protecting the identity
of the members performing the transactions from the validators and the public. The private key is sent confidentially by
the manufacturer to the distributor, with the public key that was shared earlier. The distributor uses both keys to decrypt
the encoded message, thereby completing the transaction. All transactions performed are recorded in a distributed ledger
that can be viewed by stakeholders of the supply chain. The emissions data is collected and recorded based on how it was
 International Journal of Production Research 2231

generated (during transportation or manufacturing, or while goods are being held in inventory), with the collected data then
being transferred along the movement of both goods and associated carbon assets. The average value of emissions can be
calculated per product, which can be viewed in the distributed ledger with respect to its batch number. The status of an outlet
can be evaluated using this data inherent in smart contracts.

4.3. The working of smart contracts

Smart contracts automatically enforce a given set of instructions to control emissions by suggesting possible changes.
Instructions notify supply chain managers when the threshold limits of emissions are out of bounds; these triggers are based
on the blockchain-based algorithm. Changes can include adjustments to inventory and the reduction of working hours in
specific manufacturing plants. The collected emissions data of the supply chain is added directly to the blocks, forming a
chain containing ledgers. An upper limit (UL) and a lower limit (LL) is set for each smart contract. If the emission rate
observed is beyond the upper limit or below the lower limit, then the smart contract triggers modifications in the supply
chain. As such, the mechanism displays the current status of the system to the manager for optimisation. If the emission
surpasses the upper limit, the system issues a notification that the emission from that particular outlet is higher. This then
results in an increase in the ‘cost incurred due to emissions’ (computed by multiplying the emission quantity by the tax
per unit). The obtained data can then be used for minimising the total cost by calculating the economic emissions quantity
for each outlet. The observed data being below the lower limit implies that the emissions in the respective outlet are less.
In this scenario, the proposed algorithm will change the supply of goods at the particular outlet in accordance with the
workload and the emissions in the other outlets in order to obtain a balanced system. Optimising the total emissions triggers
an increment in system efficiency. Data is added in the blocks, which are linked to form a chain, after the miners have
performed the validation procedure. By using our proposed blockchain framework within a multi-echelon supply chain and
our stated blockchain-based algorithm, we can monitor the rise and fall of carbon emissions, thereby reducing the total cost.

5. The blockchain-based approach

With this blockchain framework as a foundation, the blockchain-based supply chain approach is developed to improve the
accessibility of the network by simultaneously improving the efficiency of the system. The process is divided into three
phases: initialising, intervening and optimising.

5.1. Initialising phase

The blockchain process begins with the initialising phase where all data variables of the multi-echelon supply chain are
determined. Figure 3 explains the initialisation in a stepwise fashion. The implementation of blockchain-based smart
contracts for optimising a multi-echelon production distribution supply chain is described next.
Step 1: Initialisation of data variables for the multi-echelon supply chain (e.g. ordering cost, inventory cost, plant
location, demand).
Step 2: Presentation of the design to the decision makers for the multi-echelon supply chain.
Step 3: Preprocessing calculates the EOQ for each outlet, the inventory storage for each element of the supply chain,
and the initial optimal costs. This is done by considering storage, transportation, ordering and manufacturing costs in
regular and overtime production with safety stock, and their respective emissions.
Step 4: Smart contracts are introduced to link the supply chain to the blockchain. Smart contracts are initialised at each
outlet of the supply chain, with the smart contracts having their own threshold limits adjusted within them.
Smart contracts are initiated to record the carbon emissions, perform the emission transactions of carbon assets in the
direction of the goods transfer, and share the data in a distributed ledger. Through these transactions, the emissions data is
recorded and stored in the distributed ledger. These settings complete the initialising phase.

5.2. Intervening phase

The power of blockchain is leveraged in this phase, with the carbon emission quantities being shared in terms of their
carbon assets, following a decentralised approach to form a distributed ledger without any third-party involvement. The
carbon assets (which are carbon currency, each of a certain carbon emission quantity) are transacted by using smart contracts
following the depicted framework. The transactions, which are continuous and automated, are hashed and can be viewed by
the manager. The system shares the transaction value as the emissions amount is released due to the particular dispatched
product.
2232 V. K. Manupati et al.

Figure 3. Flowchart for the initialisation.

The emissions data records are posted in the distributed ledger, which can be viewed by the stakeholders in the supply
chain. The smart contract monitors for an increase in the emission rate at each outlet. If the emission rate is found to exceed
the threshold limit for a particular outlet involving any such transaction, the smart contract triggers the main operating
system to analyse the supply chain with the new emission data generated at that moment.
At this point, the adaptive algorithm is adjusted with updated emission data and discloses the operating demands for
each outlet to calculate the total cost of the present supply chain. This cyclical process of data sharing and optimising is
done in iterations until an optimal solution is found. Once this has been accomplished, the system stops and prepares for the
next trigger. Figure 4 shows a stepwise explanation of the intervening phase.

Step 1: Consider carbon emissions as carbon asset coins for transaction purposes.
Step 2: Set the threshold limit for the asset coins at each outlet of the supply chain.
Step 3: The supply chain’s carbon emission data is obtained through smart contracts.
Step 4: The emission rate is calculated by dividing the total emissions by the order quantity of each outlet. This provides
the emission rate for that transaction.
Step 5: When a transaction occurs in the blockchain, a loop is set up with an IF-ELSE condition such that, if any
outlet crosses the threshold limit of the carbon assets, the smart contract triggers and activates the algorithm (Step 6–
Step 8).
Step 6: If emission data is out of bounds, the smart contracts trigger and inform the system to perform further
optimisation.
Step 7: On analysing the carbon emission data from each outlet through the help of smart contracts, the data sets are
retained for further use as inputs in the optimising phase.
Step 8: The collected information is uploaded into the decentralised distributed ledger by the smart contracts.

This loop, which checks whether the process is optimised, proceeds until the stated conditions are satisfied. If there is a
chance for further optimisation beyond the regular level, it can be done accordingly. As such, with the newly retrieved data,
we can calculate the total cost with the inclusion of carbon tax. After several iterations, the optimising phase should begin.
Figure 2 presents a detailed explanation of the blockchain-based transaction of carbon assets in the supply chain.

5.3. Optimising phase

The optimising phase uses the data retrieved from the distributed ledger-based blockchain approach, as described in section
5.2. Figure 5 illustrates the blockchain-based distributed ledger process for retrieving and analysing the data for the final
optimisation, the phase concerned with minimising both total cost and carbon emissions.
 International Journal of Production Research 2233

Figure 4. Flowchart for the intervening phase.

Step 1: By using the proposed blockchain model, carbon emission data is retrieved with the help of smart contracts. In
analysing the updated carbon emission rates from the distributed ledger at individual outlets, we can store the retrieved
data and use it as an input for initialisation of the optimising phase.
Step 2: The final optimisation of the initialised data and the carbon assets (from the distributed ledger) is performed by
solving the complex MINLP problem to derive the total cost for every iteration. This optimised data, along with all
previous data can be viewed in the form of a distributed ledger in terms of transactions. An optimised system with
minimal total cost and least possible carbon emissions is thereby obtained.
Step 3: Miners are rewarded with a portion of the profit, which is determined by the validation of the transactions needed
for the optimisation. This initiative can generate interest among external participants to carry the transactions forward,
and contribute to preserving the environment.

6. Experimentation
The proposed approach is designed and executed using MATLAB software. This section presents an illustrative example
to demonstrate how the decision variables and objective functions are constructed. A perfectly functioning system is rare,
but there are methods, such as error proofing mechanisms, that can help work towards an ideal system. Accordingly, in our
proposed approach, smart contracts can set off disturbances at any of the units in the supply chain, but these disturbances
2234 V. K. Manupati et al.

Figure 5. Flowchart for the optimisation.

can then be rectified. When elevated carbon emissions in a particular unit are found, the workload at the respective outlets
can be rebalanced to keep the emissions at acceptable levels.
Traditional processes are generally incapable of continuously monitoring any disturbances and any complex carbon
asset transfers in the supply chain. We therefore proposed leveraging the distributed ledger-based blockchain approach
for adept decision making and enhanced negotiation ability between modules. This distinguishes our model from those in
the literature. Figure 1 outlines the multi-echelon supply chain with two suppliers, three manufacturing plants, and four
warehouses. Table 1 delineates the constraints, such as the demand ranging from 200 to 600, variance ranging from 10 and

Table 1. Constraint ranges.

Constraint Range values
Demand [200, 250, 300, . . . 600] units
Demand variance [10, 15, 20, . . . 40] percent
Lead time [4, 5, 6, . . . 10] days
Carbon tax [25, 30, 35, . . . 50] units

Table 2. Parameters influencing the manufacturing plant.

Plant M1 M2 M3
Fixed cost of the plant (FM ) 20 40 50
Cost of regular time production (CMPT ) 2 4 3.5
Cost of over-time production (CWPMt ) 3 6 7.5
Fixed emissions due to manufacturing (EMFMt ) 30 40 50
Variable emissions due to manufacturing (EMVMt ) 3 4 5

Table 3. Parameters influencing the supplier and the manufacturing plant.

Fixed Emissions Variable Emissions
Transportation due to due to
Unit Inventory Cost (from Transportation Transportation Emission due to
Holding Cost Ordering Cost Supplier to (from Supplier to (from Supplier to Inventory at
(HMI ) (OMI ) Plant) (TMSI ) Plant) (TEFISt ) Plant) (TEVISt ) Plants EIIMt
M1 M2 M3 M1 M2 M3 M1 M2 M3 M1 M2 M3 M1 M2 M3 M1 M2 M3
I1 0.3 0.5 – 25 30 – 0.2 0.1 – 40 35 – 2.5 3 – 1.5 1.2 –
I2 – 0.4 0.3 – 35 20 – 0.3 0.3 – 50 45 – 4 3.5 – 1.3 1.4
 International Journal of Production Research 2235

Table 4. Parameters influencing the distributor and manufacturing plant.

Fixed Emissions Variable Emissions
due to due to
Transportation Transportation Transportation Emission due to
Unit Inventory Cost (from (from Plant to (from Plant to Inventory at
Holding Cost Supplier to Warehouse) Warehouse) Warehouse
(HWP ) Plant) (TWMP ) (TEFPMt ) (TEVPMt ) (EIWPt )
W1 W2 W3 W4 W1 W2 W3 W4 W1 W2 W3 W4 W1 W2 W3 W4 W1 W2 W3 W4
P1 0.3 – 0.4 – 0.1 – 0.3 – 20 – 30 – 1.5 – 1 – 1.3 – 1.4 –
P2 – 0.2 – – – 0.2 – – – 40 – – – 2 – – – 1.2 – –
P3 – – 0.3 0.5 – – 0.2 0.1 – – 45 50 – – 3 2.5 – – 1.3 1.5

40, acquisition lead times of 4–10 days, and carbon tax expenses ranging from 25 to 50. The service level is set across
the supply chain at 1.95, where I signifies the amount of raw materials sent by supplier facility S, which is required by
manufacturing facility M to produce a unit of product P satisfying demand D of the distribution centre W.
Table 2 presents the parameters influencing the manufacturing plant, including fixed cost, regular and overtime pro-
duction cost, and emissions by manufacturing at each individual unit. While the cost indicates the expense, the emission
quantity reflects the number of carbon credits utilised. Parameters influencing the shipment of goods from the supplier to
the manufacturing plant and from the manufacturing plant to the distributors are inventory holding cost, order cost, trans-
portation cost, fixed emissions due to transportation, and emissions due to inventory; these values are provided in Tables 3

(a) (b)

(c) (d)

Figure 6. Mean Demand, Demand Variance, Lead Time and Carbon Tax vs. Total Operational Cost.
2236 V. K. Manupati et al.

and 4. The proposed approach is beneficial for the effective negotiation between the outlets by continuously monitoring the
multi-echelon supply chain using the distributed ledger mechanism of the blockchain by linking it with smart contracts.

7. Results and discussion

To obtain the most optimal solution to this problem, the results are plotted based on several considerations. For the proposed
distributed ledger-based blockchain model to be deemed superior, the total cost is computed by varying the constraints and
plotting graphs that compare the solution with that obtained through the NSGA-II.
Figure 6(a) depicts the relationship between mean demand (incremented by 50) and total operational cost; lead time
and carbon tax are again held constant at 7 days (with a variance of 35) and 35 unit costs, respectively. When comparing the
NSGA-II to the proposed blockchain approach, the results of the latter are on average 2.54% better. Table 5 offers details
on these comparisons under various levels of mean demand.
Figure 6(b) plots the total operational cost against a change in variance in 5-unit increments; mean demand, lead
time and carbon tax are held constant at 400, 7 and 35 units, respectively. When the two approaches are compared, the
blockchain-based approach yields an average of 2.96% more total cost reduction than the NSGA-II approach, demonstrating
the superiority of our proposed model under various system changes. Table 6 offers details on these results.
Figure 6(c) shows the relationship between lead time (ranging from 4 to 10, in 1-unit increments) and the corresponding
operational cost. Carbon tax is held constant at 40 and mean demand is held constant at 500. The figure illustrates the com-
parison between the blockchain-based approach and NSGA-II, demonstrating the superiority of the former. Specifically, the
total cost for the blockchain approach was approximately 1.84% less than the cost incurred with NSGA-II. Accompanying
Table 7 shows the results of this comparison along the seven lead time increments.

Table 5. Mean demand comparison table.

Mean Demand NSGA-II Blockchain Approach
200 5381.18 5254.33
250 6566.74 6406.10
300 7746.29 7551.65
350 8921.35 8692.48
400 10092.72 9829.58
450 11261.15 10963.62
500 12427.13 12095.10
550 13591.02 13224.40
600 14753.08 14351.81

Table 6. Demand variance comparison table.

Demand Variance NSGA-II Blockchain Approach
10 9771.11 9760.94
15 9788.82 9778.66
20 9803.75 9793.59
25 9816.91 9806.75
30 9828.80 9818.64
35 9839.74 9829.58
40 9849.92 9839.76

Table 7. Lead time comparison table.

Lead Time NSGA-II Blockchain Approach
4 12491.90 12267.98
5 12506.28 12281.36
6 12518.37 12293.46
7 12529.50 12304.58
8 12540.85 12314.94
9 12551.58 12324.66
10 12560.77 12333.86
 International Journal of Production Research 2237

Table 8. Carbon tax comparison table.

Carbon Tax NSGA-II Blockchain Approach
25 9734.53 9490.17
30 9910.13 9659.90
35 10085.73 9829.58
40 10261.32 9999.19
45 10436.92 10168.75
50 10612.52 10338.25

Figure 6(d) shows the relationship between carbon tax in 5-unit increments, and total operational cost; lead time is held
constant at seven days and mean demand at 400 units. The proposed blockchain approach is again superior by an average
of 2.58%. This illustrates the significant potential to reduce carbon emissions using blockchain technology. Table 8 offers
detailed values for this comparison.
The proposed blockchain-based approach outperformed the NSGA-II alternative, which focuses on minimising oper-
ational expenses, with the total optimal cost representing a summation of the operational cost and the cost incurred by
emissions. As such, the NSGA-II model solves the problem by considering two objective functions, one based on the min-
imisation of the operational expenses and the other one based on the computation of carbon emissions adhering to the
prior objective. This makes the NSGA-II model more efficient than other approaches. However, as technology is becom-
ing more advanced, most notably through the possibilities of blockchain technology, new optimal solutions are feasible, as
demonstrated in this article.
Let us consider a scenario in which the value of the total optimal cost obtained by the blockchain-based method is lower
than the total optimal cost obtained by the NSGA-II approach under similar conditions. While the NSGA-II algorithm might
produce a lower value in terms of operational costs (only manufacturing, transportation and inventory) in comparison to
the blockchain-based method, due to the nature of the algorithm to minimise multiple equations, when it comes to the total
optimal cost (which includes costs incurred due to emissions) the blockchain approach yields better results. Therefore, the
cost incurred due to emissions of the supply chain is the decisive factor for this reduction in the total optimal cost of the
supply chain via the blockchain method. There is therefore no possibility of obtaining a lower total optimal cost unless the
cost incurred due to emissions is reduced in the system. As such, as shown in the analytics presented, the total operational
cost for the proposed blockchain is lower than in the NSGA-II approach. It can therefore be concluded with confidence
that the blockchain approach enables more environmentally conscious decisions. In addition, our approach enables the dual
consideration of both operational cost and the cost incurred due to emissions, synchronising the operational expenses and
carbon emissions, and balancing economic and environmental conditions.

8. Conclusion
Advancements in technology, such as real-time monitoring systems, have advanced supply chain optimisation opportunities.
This is especially true for blockchain technology, which is destined to revolutionise supply chain management due to its
ability to enhance visibility and traceability. These features have become increasingly important, due to governmental and
regulatory policies focused on the environment, and especially on carbon emissions. Within this context, we developed a
blockchain-based approach for monitoring supply chain performance and optimising both emission levels and operational
costs in a synchronised fashion, yielding the optimal outcome for the supply chain. This approach facilitates the efficient and
effective flow of data across the supply chain while ensuring the security of the system. As such, we developed a distributed
ledger-based blockchain approach for different production allocation problems within a multi-echelon supply chain under a
carbon taxation policy. The approach integrated recent advances in digitalisation of operations with increasingly stringent
regulatory environmental policies, rendering our work both critically important and practically relevant. We simultaneously
considered production, distribution and inventory control decisions in a production allocation-based MESC problem with
lead time considerations under emission rate constraints (imposed by a carbon taxation policy). We formulated the problem
as MINLP model, results of which support the robust performance of the proposed approach. The feasibility of the approach
was validated by a comparison of the results to a non-dominated sorting genetic algorithm (NSGA-II), offering confidence
to policymakers and supply chain executives for the application of this approach.
What made our approach unique is that it integrated lead time constraints, which prior work emphasised as critical (e.g.
Banasik et al. 2018; Hammami and Frein 2014; Hammami, Nouira, and Frein 2015). In addition, our approach is closer to
reality than prior investigations in this area, since we also consider raw materials and intermediate products when computing
final product emission values as part of a multi-echelon investigation. As such, decision makers can rely on our approach to
2238 V. K. Manupati et al.

aid them in their optimisation efforts. We also contribute to the area of carbon footprint considerations in the supply chain,
which has been scarce in the operations management domain (Benjaafar, Li, and Daskin 2013). In addition, research noted
the challenges associated with optimising and simulating sustainable supply chains to attain ideal solutions (e.g. Hammami
and Frein 2014), and to our best knowledge, no research has considered the dual, synchronous optimisation of both carbon
emissions and operational costs. The problem was enhanced by the ability to consider the impact of both lead time and safety
stock constraints, as well as differences in regular and overtime production rates. Another distinctive aspect of our work is
that we leveraged the power of blockchain technology, allowing the optimisation of a multi-echelon supply chain via live
data inspection and traceability. While the literature has covered multi-variable, multi-echelon supply chain optimisation,
we have found no studies that performed a live monitoring of the supply chain within the context of smart contracts.
Interoperability has begun to recognise the potential of new technologies to drive a holistic view of the value chain,
towards a robust chain of integrated elements. In this context, blockchain is a credible platform characterised by traceability
among supply chain members. This was a formidable way to tackle carbon emission concerns, which affect all links in the
supply chain. As such, buyers and sellers can exchange carbon allowances or consumption units efficiently and effectively,
along with the needed transparency and traceability. Carbon emissions can be reported, in compliance with the blockchain
traceability mechanisms, and compliance to carbon taxation policy can be ensured. At the same time, consumers can assess
product standards and the environmental concerns exhibited by the players in the supply chain. The potential is consid-
erable, offering insight and tracking ability for a multitude of performance metrics, including lead times, transportation,
distribution, and storage. It can also help manufacturers and consumers comply with government regulations, drive prod-
uct innovation and inform policy formulation. Taken together, blockchain-based supply chains can enable firms to build
consortia to improve connected supply chain operations.
As such, our work has the potential for adoption by other sectors, including industrial manufacturing, hospitality
services, government services, railways, and educational services. For instance, blockchain technology coupled with appro-
priate algorithms can significantly improve the implementation of health care services, due to its ability to manage patient
data and prescriptions. Moreover, such platforms can direct patients to take the correct medicine at the right dosages.
The incorporation of blockchain technology into applications can thus improve user safety. The potential applications are
endless. We hope that this article offers inspiration and motivation for continued advancements in this area.

Acknowledgements
We would also like to thank the special issue guest editors, the associate editor and the four anonymous reviewers for their constructive
feedback on earlier versions that helped us to improve the paper.

Disclosure statement
No potential conflict of interest was reported by the authors.

ORCID
Stephan M. Wagner http://orcid.org/0000-0003-0471-5663

References
Arampantzi, C., and I. Minis. 2017. “A New Model for Designing Sustainable Supply Chain Networks and Its Application to a Global
Manufacturer.” Journal of Cleaner Production 156: 276–292.
Babazadeh, R., J. Razmi, M. S. Pishvaee, and M. Rabbani. 2017. “A Sustainable Second-Generation Biodiesel Supply Chain Network
Design Problem Under Risk.” Omega 66: 258–277.
Banasik, A., J. M. Bloemhof-Ruwaard, A. Kanellopoulos, G. D. H. Claassen, and J. G. van der Vorst. 2018. “Multi-criteria Decision
Making Approaches for Green Supply Chains: A Review.” Flexible Services and Manufacturing Journal 30 (3): 366–396.
Benjaafar, S., Y. Li, and M. Daskin. 2013. “Carbon Footprint and the Management of Supply Chains: Insights From Simple Models.”
IEEE Transactions on Automation Science and Engineering 10 (1): 99–116.
Bode, C., and S. M. Wagner. 2015. “Structural Drivers of Upstream Supply Chain Complexity and the Frequency of Supply Chain
Disruptions.” Journal of Operations Management 36: 215–228.
Bouchery, Y., A. Ghaffari, Z. Jemai, and Y. Dallery. 2012. “Including Sustainability Criteria Into Inventory Models.” European Journal
of Operational Research 222 (2): 229–240.
Bové, A. T., and S. Swartz. 2016. “Starting at the Source: Sustainability in Supply Chains.” McKinsey & Company White Paper,
November.
 International Journal of Production Research 2239

Chaabane, A., A. Ramudhin, and M. Paquet. 2012. “Design of Sustainable Supply Chains Under the Emission Trading Scheme.”
International Journal of Production Economics 135 (1): 37–49.
Choi, T. M. 2013. “Optimal Apparel Supplier Selection with Forecast Updates Under Carbon Emission Taxation Scheme.” Computers &
Operations Research 40 (11): 2646–2655.
Cole, R., J. Aitken, and M. Stevenson. 2019. “Blockchain Technology: Implications for Operations and Supply Chain Management.”
Supply Chain Management: An International Journal 24 (4): 469–483.
Crosby, M., P. Pattanayak, S. Verma, and V. Kalyanaraman. 2016. “Blockchain Technology: Beyond Bitcoin.” Applied Innovation Review
2: 6–9.
Devika, K., A. Jafarian, and V. Nourbakhsh. 2014. “Designing a Sustainable Closed-Loop Supply Chain Network Based on Triple Bottom
Line Approach: A Comparison of Metaheuristics Hybridization Techniques.” European Journal of Operational Research 235 (3):
594–615.
Elhedhli, S., and R. Merrick. 2012. “Green Supply Chain Network Design to Reduce Carbon Emissions.” Transportation Research Part
D: Transport and Environment 17 (5): 370–379.
Eskandarpour, M., P. Dejax, J. Miemczyk, and O. Péton. 2015. “Sustainable Supply Chain Network Design: An Optimization-Oriented
Review.” Omega 54: 11–32.
Fahimnia, B., J. Sarkis, A. Choudhary, and A. Eshragh. 2015. “Tactical Supply Chain Planning Under a Carbon Tax Policy Scheme: A
Case Study.” International Journal of Production Economics 164: 206–215.
Fard, A. M. F., and M. Hajaghaei-Keshteli. 2018. “A Tri-Level Location-Allocation Model for Forward/Reverse Supply Chain.” Applied
Soft Computing 62: 328–346.
Fu, B., Z. Shu, and X. Liu. 2018. “Blockchain Enhanced Emission Trading Framework in Fashion Apparel Manufacturing Industry.”
Sustainability 10 (4): 1105.
Galvez, J. F., J. C. Mejuto, and J. Simal-Gandara. 2018. “Future Challenges on the Use of Blockchain for Food Traceability Analysis.”
TrAC Trends in Analytical Chemistry 107: 222–232.
Garcia-Torres, S., L. Albareda, M. Rey-Garcia, and S. Seuring. 2019. “Traceability for Sustainability – Literature Review and Conceptual
Framework.” Supply Chain Management: An International Journal 24 (1): 85–106.
Hammami, R., and Y. Frein. 2014. “A Capacitated Multi-Echelon Inventory Placement Model Under Lead Time Constraints.” Production
and Operations Management 23 (3): 446–462.
Hammami, R., I. Nouira, and Y. Frein. 2015. “Carbon Emissions in a Multi-Echelon Production-Inventory Model with Lead Time
Constraints.” International Journal of Production Economics 164: 292–307.
Hua, G., T. C. E. Cheng, and S. Wang. 2011. “Managing Carbon Footprints in Inventory Management.” International Journal of
Production Economics 132 (2): 178–185.
Ivanov, D., A. Dolgui, and B. Sokolov. 2019. “The Impact of Digital Technology and Industry 4.0 on the Ripple Effect and Supply Chain
Risk Analytics.” International Journal of Production Research 57 (3): 829–846.
Jaber, M. Y., C. H. Glock, M. A. Ahmed, and E. I. Saadany. 2013. “Supply Chain Coordination with Emissions Reduction Incentives.”
International Journal of Production Research 51 (1): 69–82.
Kamble, S. S., A. Gunasekaran, and S. A. Gawankar. 2020. “Achieving Sustainable Performance in a Data-Driven Agriculture Supply
Chain: A Review for Research and Applications.” International Journal of Production Economics 219: 179–194.
Kannegiesser, M., H.-O. Günther, and N. Autenrieb. 2015. “The Time-to-Sustainability Optimization Strategy for Sustainable Supply
Network Design.” Journal of Cleaner Production 108 (Part A): 451–463.
Koh, S. L., A. Genovese, A. A. Acquaye, P. Barratt, N. Rana, J. Kuylenstierna, and D. Gibbs. 2013. “Decarbonising Product Supply
Chains: Design and Development of an Integrated Evidence-Based Decision Support System – The Supply Chain Environmental
Analysis Tool (SCEnAT).” International Journal of Production Research 51 (7): 2092–2109.
Kumar, G., N. Subramanian, and M. Ramkumar. 2018. “Missing Link Between Sustainability Collaborative Strategy and Supply Chain
Performance: Role of Dynamic Capability.” International Journal of Production Economics 203: 96–109.
Letmathe, P., and N. Balakrishnan. 2005. “Environmental Considerations on the Optimal Product Mix.” European Journal of Operational
Research 167 (2): 398–412.
Liu, Q., W. Zhang, M. Yao, and J. Yuan. 2017. “Carbon Emissions Performance Regulation for China’s Top Generation Groups by 2020:
Too Challenging to Realize?” Resources, Conservation and Recycling 122: 326–334.
Manupati, V. K., S. Deo, N. Cheikhrouhou, and M. K. Tiwari. 2012. “Optimal Process Plan Selection in Networked Based Manufacturing
Using Game-Theoretic Approach.” International Journal of Production Research 50 (18): 5239–5258.
Manupati, V. K., S. J. Jedidah, S. Gupta, A. Bhandari, and M. Ramkumar. 2019. “Optimization of a Multi-Echelon Sustainable
Production–Distribution Supply Chain System with Lead Time Consideration Under Carbon Emission Policies.” Computers &
Industrial Engineering 135: 1312–1323.
Mathiyazhagan, K., and A. N. Haq. 2013. “Analysis of the Influential Pressures for Green Supply Chain Management Adoption – An
Indian Perspective Using Interpretive Structural Modeling.” International Journal of Advanced Manufacturing Technology 68
(1-4): 817–833.
Narasimhan, R., T. Schoenherr, B. W. Jacobs, and M. K. Kim. 2015. “The Financial Impact of FSC Certification in the United States: a
Contingency Perspective.” Decision Sciences 46 (3): 527–563.
Queiroz, M. M., and S. F. Wamba. 2019. “Blockchain Adoption Challenges in Supply Chain: An Empirical Investigation of the Main
Drivers in India and the USA.” International Journal of Information Management 46: 70–82.
2240 V. K. Manupati et al.

Ramkumar, M., and M. Jenamani. 2015. “Sustainability in Supply Chain Through e-Procurement – An Assessment Framework Based on
DANP and Liberatore Score.” IEEE Systems Journal 9 (4): 1554–1564.
Saberi, S., M. Kouhizadeh, J. Sarkis, and L. Shen. 2019. “Blockchain Technology and its Relationships to Sustainable Supply Chain
Management.” International Journal of Production Research 57 (7): 2117–2135.
Samadi, A., N. Mehranfar, A. M. Fathollahi Fard, and M. Hajiaghaei-Keshteli. 2018. “Heuristic-based Metaheuristics to Address a
Sustainable Supply Chain Network Design Problem.” Journal of Industrial and Production Engineering 35 (2): 102–117.
Schmidt, C. G., and S. M. Wagner. 2019. “Blockchain Technology and Supply Chain Governance – A Transaction Cost Theory
Perspective.” Journal of Purchasing & Supply Management 25 (4): 100552.
Sharma, R., S. S. Kamble, and A. Gunasekaran. 2018. “Big GIS Analytics Framework for Agriculture Supply Chains: A Literature
Review Identifying the Current Trends and Future Perspectives.” Computers and Electronics in Agriculture 155: 103–120.
Sundarakani, B., R. De Souza, M. Goh, S. M. Wagner, and S. Manikandan. 2010. “Modeling Carbon Footprints Across the Supply Chain.”
International Journal of Production Economics 128 (1): 43–50.
Tian, F. 2016. “An Agri-Food Supply Chain Traceability System for China Based on RFID & Blockchain Technology.” In 2016 13th
International Conference on Service Systems and Service Management (ICSSSM), KUST, Kunming, China, June, edited by Baojian
Yang.
Tognetti, A., P. T. Grosse-Ruyken, and S. M. Wagner. 2015. “Green Supply Chain Network Optimization and the Trade-off Between
Environmental and Economic Objectives.” International Journal of Production Economics 170 (Part B): 385–392.
Tseng, S. C., and S. W. Hung. 2014. “A Strategic Decision-Making Model Considering the Social Costs of Carbon Dioxide Emissions for
Sustainable Supply Chain Management.” Journal of Environmental Management 133: 315–322.
Varsei, M., and S. Polyakovskiy. 2017. “Sustainable Supply Chain Network Design: A Case of the Wine Industry in Australia.” Omega
66: 236–247.
Wahab, M. I. M., S. M. H. Mamun, and P. Ongkunaruk. 2011. “EOQ Models for a Coordinated Two-Level International Supply Chain
Considering Imperfect Items and Environmental Impact.” International Journal of Production Economics 134 (1): 151–158.
Yang, L., J. Wang, and J. Shi. 2017. “Can China Meet its 2020 Economic Growth and Carbon Emissions Reduction Targets?” Journal of
Cleaner Production 142: 993–1001.
Zakeri, A., F. Dehghanian, B. Fahimnia, and J. Sarkis. 2015. “Carbon Pricing Versus Emissions Trading: A Supply Chain Planning
Perspective.” International Journal of Production Economics 164: 197–205.
Zhao, G., S. Liu, C. Lopez, H. Lu, S. Elgueta, H. Chen, and B. M. Boshkoska. 2019. “Blockchain Technology in Agri-Food Value Chain
Management: A Synthesis of Applications, Challenges and Future Research Directions.” Computers in Industry 109: 83–99.
 International Journal of Production Research 2241

Appendix

Table A1. Variable description.

Variable Description
S Supplier
M Manufacturing plant
W Distribution centre
I Item to be supplied to manufacturing plant M
P Product delivered to distribution centre W
D Demand
LMI Lead time
rMI Reorder point
Z1−α Service level of the supply chain
σ LT Variance in demand during the lead time
HMI Holding cost of item I at plant M
QMI Order quantity at plant M
OMI Ordering cost of I
ScapM Supply limit for items to manufacturing plant M
PcapMP Production limit for product P in manufacturing plant M
FM Fixed cost of operating plant M
μWM Mean demand of product P
σ WM Variance in demand of product P
μDWP Mean of customer demand at W for product P
σ DWP Variance of customer demand at W for product P
DMI Demand of item I in plant M
VMI Variance of item I in plant M
HWP Holding cost at distribution centre P
QRWPMt Regular time production quantity of product P from M to W
QOWPMt Over-time production quantity of product P from M to W
CMPt Cost per unit of regular time production of product P at M
CWPMt Cost per unit of over-time production of product P at M
TMSI Transportation cost of item I from supplier S to plant M
TWMP Transportation cost of product P from plant M to distribution centre W
EFM Fixed emissions from plant M
EVM Variable emissions from plant M
TEFPM Fixed emissions due to the transportation of product P from plant M
TEVPM Variable emissions due to the transportation of product P from plant M
TEFIS Fixed emissions due to the transportation of item I from supplier S
TEVIS Variable emissions due to the transportation of item I from supplier S
EIPW Emissions due to inventory at distribution centre W
EIPMt Emissions due to inventory at plant M for time t
T Carbon tax
Ψ Trading cost of carbon credits
XM Binary variable that is
0: Plant M is open
1: Plant M is closed
YWMP Binary variable that is
0: Material is transported to warehouse W from plant M
1: Material is transported to warehouse W from plant M
ZMSI Binary variable that is
0: Supplier S serves item I to plant M
1: Supplier does not serve plant M
S Binary variable that is
0: Smart contract is triggered
1: Smart contract is not triggered
Z Total operating cost
Copyright of International Journal of Production Research is the property of Taylor & Francis
Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv
without the copyright holder's express written permission. However, users may print,
download, or email articles for individual use.