ee UNIT-4 KNOWLEDGE REPRESENTATION AND CLASSICAL PLANNING i ONTOLOGICAL ENGINEERING ke nema as : tology 18 a brarich (Of HietapHYSTEs) NVBICH is a’ BraHCHIOE phlilosOphy that deals with studying being, existence, and feality. Ontology Engineering is “the set of activities that concern the ontology development process, the ontology life cycle, and the methodologies, tools and languages for building ontologies” jologies are a powerful tool for organizing and understanding information in a structured way. They provide a clear framework for defining the relationships between different concepts, making it easier to share and analyze data across various fields. Ontologi Ontologies are formal definitions of vocabularies that allow us to define difficult or complex structures and new relationships between vocabulary terms and members of classes that we define. Ontologies generally describe specific domains such as scientific research areas. Example: Ontology depicting Movie:-Components of Ontology: 1. Individuals - Individuals are also known as instances of objects or concepts. It may or may not be present in an ontology. It represents the atomic level of an ontology. For example, in the above ontology of movie, individuals can be a film (Titanic), a director (James Cameron), an actor (Leonardo DiCaprio). 1. Classes — Sets of collections of various objects are termed as classes. For example, in the above ontology representing movie, movie genre (e.g. Thriller, Drama), and types of person (Actor or Director) are classes. 1. Attributes — Properties that objects may possess. For example, a movie is described by the set of ‘parts’ it contains like Script, Director, and Actors. 1. Relations — Ways in which concepts are related to one another. For example, as shown above in the diagram a movie has to have a script and actors in it. Different Ontology Languages: + CycL — It was developed for the Cyc project and is based on First Order Predicate Calculus. + Rule Interchange Format (RIF) — It is the language used for combining ontologies and rules. + Open Biomedical Ontologies (OBO) — It is used for various biological and biomedical ontologies. + Web Ontology Language (OWL) - It is developed for using ontologies over the World Wide Web (WWW). Ontology Engineering Steps: 1. Domain Analysis: Identify the domain, scope, and requirements of the ontology. / 2. Knowledge Acquisition: Gather knowledge from experts, literature, and data sources. ae 3. Conceptualization: Identify key concepts, relationships, and rules. 4, Formalization: Represent the conceptualization using a formal language (e.g., OWL, RDF).Categories: 5. Implementation: Create the ontology using tools (¢.g., Protégé, TopBraid), 6. Testing and Validation: Verify the ontology's consistency, completeness, and accuracy. 7. Maintenance and Evolution: Update and refine the ontology as the domain evolves. Applications of Ontology Engineering: - Knowledge Management: Ontologies enable the representation and sharing of knowledge across organizations. . Artificial Intelligence: Ontologies provide a foundation for AI systems to reason and make decisions. . Data Integration: Ontologies facilitate the integration of data from disparate sources. 4. Natural Language Processing: Ontologies enable the representation of linguistic knowledge and support NLP applications. 5. Semantic Web: Ontologies are a key component of the Semantic Web, enabling the creation of a web of meaning. N w CATEGORIES AND OBJECTS In ontology, categories and objects are fundamental concepts used to represent knowledge and describe the world. Categories, also known as concepts or classes, are abstract representations of sets of objects that share common characteristics, properties, or attributes. Categories are the basic building blocks of an ontology and are used to organize and structure knowledge. Examples of categories: 1. Animal 2. Vehicle 3. City4, Person 5. Organization aE i ities that belong to i Iso known as instances or individuals, are specific Ree ste is category. Objects have unique identities and are cl a particular ny. properties and attributes. Examples of objects: 1. Lion (an instance of the category Animal) 2. Toyota Camry (an instance of the category Vehicle) 3. New York City (an instance of the category City) 4 John Smith (an instance of the category Person) 5. Google (an instance of the category Organization) Relationships between Categories and Objects: 1. Instantiation: An object is an instance ofa category (e.g., Lion is an instance of Animal). 2. Classification: A cate; Bory is a classifi classification of Lion), ication of an object (e.g., Animal is a 3. Inheritance: An object inherits ~ Inheritan Properties and attributes from its Category (e.g., Lion inherits the Property "has four legs" from Animal), Ontology Example: Suppose we Motorcycle. Davidson have an ontolo, gy about vehicles, with categories like Car, Truck, and We can create objects like Toyota Camry, Ford F-150, and Harley-In this ontology: ~ Car, Truck, and Motorcycle are categories, ~ Toyota Camry, Ford F-150, and Harley-Davidson are objects. ~ Toyota Camry is an instance of the category Car, ~ Caria classification of Toyota Camry, EVENTS AND MENTAL EVENTS AND MENTAL, OBJECTS Tn ontology, ¢vents, mental events, and mental obj used to represent and describe the world, Event jects are fundamental concepts Events are occurrences or happeni Physical, social, or abstract, an spatial properties. Examples of events 1. A car accident ngs that take place in the world. They can be id are often characterized by their temporal and 2. A wedding ceremony 3. A scientific experiment 4. A natural disaster (¢.g., earthquake, hurricane) Mental Events: Mental events are a type of event that 0 consciousness. They are subjective ex; feelings, perceptions, or intentions. Examples of mental events: 1. A person thinking about a problem ccurs within an individual's mind or periences that can be described as thoughts, 2. A person feeling happy or sad 3. A person perceiving a visual stimulus4. A person intending to perform an action Mental Objects: Mental objects are abstract entities that exist within an individual's mind or consciousness. They can be thoughts, concepts, ideas, or mental images. Examples of mental objects: 1. A person's concept of a chair 2. A person's mental image of a loved one 3. A person's idea of a hypothetical scenario 4. A person's thought about a mathematical concept Relationships between Events, Mental Events, and Mental Objects: 1. Causality: Events can cause mental events (e.g., a car accident causes a person to feel scared). 2. Association: Mental events can be associated with mental objects (e.g., a person thinking about a chair is associated with their concept of a chair). 3. Influence: Mental objects can influence mental events (e.g., a person's concept of a chair influences their perception of a chair). 4. Representation: Mental objects can represent events or other mental objects (e.g., a person's mental image of a loved one represents their memory of that person). Ontology Example: Suppose we have an ontology about human experiences, with events, mental events, and mental objects as key concepts. We can represent the following scenario: ~ Event: A person attends a concert. - Mental Event: The person feels happy and excited during the concert.- Mental Object: The person's concept of music and their mental image of the concert. In this ontology: - The concert is an event that causes the person to feel happy and excited (mental event). - The person's concept of music and their mental image of the concert are mental objects that influence their mental event (feeling happy and excited). - The person's mental event (feeling happy and excited) is associated with their mental objects (concept of music and mental image of the concert). REAS' iG SYSTEMS FOR CATEGORIES Reasoning: The reasoning is the mental process of deriving logical conclusion and making dictions from available knowledge, facts, and beliefs. Or we can say, pre " Tt is a general process of "Reasoning is a way to infer facts from existing data. thinking rationally, to find valid conclusions. In artificial intelligence, the reasoning is essential so that the machine can also think rationally as a human brain, and can perform like a human. Types of Reasoning In artificial intelligence, reasoning can be divided into the following categories: o Deductive reasoning o Inductive reasoning o Abductive reasoning o Common Sense Reasoning o Monotonic Reasoning o Non-monotonic Reasoning 1. Deductive reasoning:Deductive reasoning is deducing new information from logically tela information. It is the form of valid reasoning, which means the conclusion must be true when the premises are true. ted Kooy, $ argue, Deductive reasoning is a type of propositional logic in Al, and it requires Various tules and facts. It is sometimes referred to as top-down Teasoning, ang contradictory to inductive reasoning. In deductive reasoning, the truth of the premises guarantees the truth of the conclusion, Deductive Teasoning mostly starts from the general premises to the specific Conclusion, which can be explained as below example. Example: Premise-1: All the human eats veggiesPremise-2: Suresh is human. Conclusion: Suresh eats veggies. The general process of deductive Teasoning is given below: OL 2. Inductive Reasoning: Inductive reasoning is a form of reasoning to arrive at a conclusion using limited sets of facts by the process of generalization. It starts with the series of specific facts or data and reaches to a general statement or conclusion. Inductive reasoning is a type of propositional logic, which is also known as cause- effect reasoning or bottom-up reasoning. In inductive reasoning, we use historical data or various premises to generate a generic rule, for which premises support the conclusion. i ion, so In inductive reasoning, premises provide probable supports to the conelusior the truth of premises does not guarantee the truth of the conclusion.Oe Example: Premise: All of the pigeons we have seen in the zoo are white. Conclusion: Therefore, we can expect all the pigeons to be white. Observations Pera -@®- @® 3. Abductive reasoning: Abductive reasoning is a form of logical reasoning which starts with single or multiple observations then seeks to find the most likely explanation or conclusion for the observation. Abductive reasoning is an extension of deductive reasoning, but in abductive reasoning, the premises do not guarantee the conclusion. Example: Implication: Cricket ground is wet if it is raining Axiom: Cricket ground is wet. Conclusion It is raining. 4. Common Sense Reasoning Common sense reasoning is an informal form of reasoning, which can be gained through experiences. Common Sense reasoning simulates the human ability to make presumptions about events which occurs on every day. It relies on good judgment rather than exact logic and operates on heuristic knowledge and heuristic rules. Example: 1. One person can be at one place at a time. 2. If I put my hand in a fire, then it will burn,Sr ee The above two statements are the examples of common sense reasoning which , human mind can easily understand and assume. 5. Monoto: Reasoning: att 7 In monotonic teasoning, once the conclusion is taken, then it will remain the same even if we add some other information to existing information in our knowledge base. In Monotonic reasoning, adding knowledge does not decrease the set of Prepositions that can be derived, To solve monotonic Problems, we can derive the valid conclusion from the available facts only, and it will not be affected by new facts, Monotonic Teasoning is used in conventional Teasoning systems, and a logic-based system is monotonic, Any theorem Proving is an example of monotonic Teasoning. Example: e Earth revolves around the Sun, It is a true fact, and it cannot be changed even if we add another sentence in e, "Th knowledge base lik ‘The moon revolves around the earth" Or "Earth is not round," ete, Advantages of Monotonic Reasoning: © In monotonic Teasoning, each old Proof will always remain valid. I We cannot Tepresent the real world scenarios using Monotonic reasoning. : Hypothesis knowledge cannot be expressed with monotonic reasoning, which means facts should be true. Sea o Since we can only derive conclusions from the old proofs, so knowledge from the real world cannot be added. 6. Non-monotonic Reasoningae In Non-monotonic reasoning, some conclusions may be invalidated if we add some more information to our knowledge base. Logic will be said as non-monotonic if some conclusions can be invalidated by adding more knowledge into our knowledge base. Non-monotonic reasoning deals with incomplete and uncertain models. “Human perceptions for various things in daily life, "is a general example of non- monotonic reasoning. Example: Let suppose the knowledge base contains the following knowledge: © Birds can fly o Penguins cannot fly o Pitty is a bird So from the above sentences, we can conclude that Pitty can fly. However, if we add one another sentence into knowledge base "Pitty is a penguin", which concludes "Pitty cannot fly", so it invalidates the above conclusion. Advantages of Non-monotonic reasoning: For real-world systems such as Robot navigation, we can use non-monotonic reasoning. In Non-monotonic reasoning, we can choose probabilistic facts or can make ° assumptions. Disadvantages of Non-monotonic Reasoning: e Innon-monotonic reasoning, the old facts may be invalidated by adding new sentences. o Itcannot be used for theorem proving. Reasoning with Default Information in AIi er Default reasoning is a fundamental concept in Artificial Intelligence (At enables systems to make assumptions and draw conclusions based on Incomplete Or uncertain information, Types of Default Reasoning: |. Closed-World Assumption (CWA): Assumes that any information not explicitly stated is false, 2. Open-World Assumption (OWA): Assumes that any information not explicitly stated may be true or false, >. Default Logic: A formal system for reasoning with defaults, introduced by Raymond Reiter. Default Reasoning Techniques: 1. Default Rules: Rules that apply unless there is explicit information to the contrary. Example: "By default, students are enrolled in the morning session." 2. Default Assumptions: Assumptions made in the absence of explicit information. Example: “Assuming the weather is sunny, we will have a picnic." 3. Circumscription: A technique for minimizing the extent of a predicate. Example: "All students are enrolled in either the morning or afternoon session." Applications of Default Reasoning: i is isions based on 1. Expert Systems: Default reasoning is used to make decision: incomplete or uncertain information. oe fea 2. Natural Language Processing (NLP): Default reasoning is us ambiguities and make inferences in text.3. Robotics: Default reasoning is used to make decisions based on incomplete or uncertain sensor data. Challenges and Limitations: 1. Inconsistent Defaults: Conflicting default rules or assumptions can lead to inconsistent conclusions. 2. Default Overriding: Defaults can be overridden by explicit information, leading to inconsistent conclusions. 3. Computational Complexity: Default reasoning can be computationally expensive, especially in large knowledge bases. CLASSICAL PLANNING planning In Artificial Intelligence (AI), planning refers to the process of finding a sequence of actions to achieve a specific goal or set of goals. Types of Planning: 1. Classical Planning: Planning with complete knowledge and deterministic actions. 2. Planning under Uncertainty: Planning with incomplete knowledge or ‘ic outcomes. probabil 3. Planning with Incomplete Knowledge: Planning with incomplete knowledge of the environment or actions. 4, Hybrid Planning: Planning that combines different planning approaches or techniques. Classical planning 1. Deterministic actions: Actions have deterministic effects, meaning that the outcome of an action is always known.2. Static environment: The environment is static, meaning that it does not Chang, unless an action is taken. \ i ete knowledge of the © 3. Complete knowledge: The planning system has comp! Be Of the environment and the actions. 4. Sequential actions: Actions are taken sequentially, one at a time. Classical Planning Techniques Classical planning stands for the assumption of a static world, where the transition between the states is deterministic, and the observable environment is fully observable. The purpose is to search for a series of actions (i.¢., a plan) which will take the current state and move it until the goal state is reached, while satisfying the given conditions and limitations. Classical planning algorithms can be broadly categorized into two main approaches: State-SpaceSearch: These algorithms make use of the state space by producing and concerning the successive states recurrently until reaching a goal state. For instance, there are breadth-first search, depth-first search and more as A* and greedy best-first search. The part of the state-space search collective is usually maintained with a frontier unexplored state and then these states systematically expanded until a solution is found or the space in the search is exhausted. Plan-SpaceSearch: Such algorithms are applied in the operational plan space, where the emphasis is on developing and fine-tuning the partial maps. Segments mentioned herein include partial-order planning, hierarchical task network planning, and search algorithms like UCPOP and VHPOP that can search the plan space. Form-space search algorithms work by continuously growing and changing these partial plans, using the constraints and dealing with the inconsistencies until there is a complete and consistent plan. ALGORITHMS FOR PLANNING WITH STATE SPACE SEARCH State space search State space search is a fundamental technique in artificial intelligence (AI) solving planning problems. In AI planning, the goal is to determine a sete actions that transitions from an initial state to a desired goal state. State sp: forclassical planning Techniques Vee abled een ace search © erate Sp ace search © par SP state SPACE goarch Pt ———_—— ) Lor ward state space search r > Backward state Space Seareh 3) Heuristic skate Spoce coarch ® forward | grate space search: faartel arate} Ve an \ Kexions| Ao \ las me a [es state) & ae Me problem se Pha pe oereplared. * RB- wo oar ports . oe Sy 7 | Poh ve A | @,» Action CPP? | CP) prctton Ceo ih) 6) —> ae Gas ? Pp Eniviolly Pa ok A ANE moved fo “B! i ‘ce search” srote Space 20" L Q wek wod| | | L | Toittal state: e8 L oe A chon (P88) Ko] GC @ ee Actton( Pp en) | : © Heuristic state Fale Space sear | ey tO waht ch node This Search Used © y. explore Siesk .’ Tha ‘we WII wveduce tme + Searching. Exomple hee oe tad and eak Calke Loo problem. Inital state - | Lave Crane) deal state — Save Cake) eotCeake) ReSgercAch'on = eak Ceale) ele a —) .— Have (coke) +00 ] Bale Cealce) | effect , e Have (cake)search algorithms explore all possible states and actions to find an optimal or feasible solution. This approach is crucial in various applications such as robotics, game playing, logistics, and scheduling. State Space Search Algorithms 1. Breadth-First Search (BFS) Breadth-First Search (BFS) explores the state space level by level, starting from the initial state. It systematically explores all possible states at each level before moving to the next level. Advantages Guarantees finding the shortest path to the goal if one exists. + Simple to implement and understand. Disadvantages Can be memory-intensive, as it needs to store all states at the current level. Applications « Puzzle solving (e.g., Rubik's Cube, sliding puzzles). « Finding shortest paths in unweighted graphs. 2. Depth-First Search (DFS) Depth-First_Search (DFS) explores as far down a branch as possible before backtracking. It uses a stack data structure to keep track of the states to be explored. Advantages Requires less memory compared to BFS. « Can be more efficient for problems with deep solutions. Disadvantages + May get stuck in deep or infinite branches. « Does not guarantee the shortest path to reach the goal. Applications + Solving mazes. + Pathfinding in games. 3. Iterative Deepening Search (IDSIterative Deepening Search IDS) combines BFS and DFS. It performs a Series ¢ d ‘easing the depth limit with each iteration, lepth-limited searches, incr Advantages + Uses less memory than BFS. + Guarantees finding the shortest path to the goal. Disadvantages + Can be slower than BFS due to repeated exploration of states. Applications . Situations where the depth of the solution is unknown. + Alin games and puzzles. 4. A* Search A* search uses a heuristic function to estimate the cost from the current state to the goal State. It explores states based on the sum of the cost to reach the state and the estimated cost to the goal (f(n) = g(n) + h(n)). Advantages + Efficiently finds the shortest path if the heuristic is admissible (never overestimates the cost). + Can handle a wide range of problems. Disadvantages + The performance depends on the quality of the heuristic. + Can be memory-intensive. Applications + Pathfinding in navigation systems. + Robotics and motion planning. 5. Greedy Best-First Search Greedy Best-First Search selects the state that appears to be closest to the ‘goal according to a heuristic function. It focuses on exploring the most promising states first. Advantages Can be faster than A* in some cases, - Simple to implement. Disadvantages Does not guarantee finding the shortest path. Can get stuck in local minima.Applications + Approximate solutions in large search spaces. + Real-time pathfinding in games. 6. Dynamic Programming Dynamic programming breaks the problem into smaller subproblems and solves each sub problem only once, storing the results. It uses these stored results to construct the solution to the larger problem. Advantages + Efficient for problems with overlapping sub problems. + Guarantees optimal solutions. Disadvantages + Requires significant memory to store intermediate results. + May not be applicable to all state space problems. Applications + Route planning in transportation networks. + Resource allocation problems. Applications of State Space Search Algorithms - Robotics: State space search algorithms enable robots to plan paths, avoid obstacles, and perform tasks autonomously in dynamic environments. . Game Playing: Algorithms like A* and BFS are used in game AI to find optimal paths, solve puzzles, and develop strategies. - Logistics and Supply Chain: State space search helps in optimizing delivery routes, warehouse management, and resource allocation. - Natural Language Processing: Dynamic programming algorithms like the Viterbi algorithm are used for tasks such as speech recognition and part-of- speech tagging. Challenges in State Space Search 1. Scalability: As the state space grows, the computational resources required increase exponentially. Efficient heuristics and pruning techniques are essential to manage large state spaces. 1. Heuristic Design: The quality of heuristic functions greatly impacts the performance of algorithms like A*. Designing effective heuristics can be challenging and problem-specific. : Memory Management: State space search algorithms can be memory- intensive, especially for large or complex problems. Techniques like iterative deepening help mitigate memory usage but may increase computation time.Eh LL PLANNING GRAPHS rily used in automated planning ang A Planning Graph is a data structure prima artificial intelligence to find solutions to planning problems. It Tepresents ; planning problem's progression through a series of levels that describe states of the world and the actions that can be taken. COMPONENTS: pes of levels: action levels and 1. Levels: A Planning graph has two alternating ty} state levels, The first level is always a state level, representing the initial state of the planning problem. J. State Levels: These levels consist of nodes representing logical propositions or facts about the world. Each successive state level contains all the propositions of the previous level plus any that can be derived by the actions of the intervening action levels. %. Action Levels: These levels contain nodes representing actions. An action node connects to a state level if the state contains all the preconditions necessary for that action. Actions in turn can create new state conditions, influencing the subsequent state level. L, Edges: The graph has two types of edges: one connecting state nodes to action nodes (indicating that the state meets the preconditions for the action), and another connecting action nodes to state nodes (indicating the effects of the action). { Mutual Exclusion (Mutex) Relationships: At each level, certain pairs of actions or states might be mutually exclusive, meaning they cannot coexist or occur together due to conflicting conditions or effects. These mutex relationships are critical for reducing the complexity of the planning problem by limiting the combinations of actions and states that need to be considered. Levels in Planning Graphs 1 state of the planning graph that consists of nodes Level SO: It is the initial each representing the state or conditions that can be true. les that are responsible for taking all Level AO: Level AO consists of nod of the initial condition described in the SO. specific actions in terms : : «Si: It represents the state or condition which could hold at a time i, it may be both P and ~P. i 2 > Ak It contains the actions that could have their preconditions satisfied at i. Working of Planning GraphThe planning graph has a single proposition level that contains all the initial conditions. The planning graph runs in stages, each stage and its key workings are described below: ie Extending the Planning Graph: At stage i (the current level), the graph plan takes the planning graph from stage i-1 (the previous stage) and extends it by one time step. This adds the next action level representing all possible actions given the propositions (states) in the previous level, followed by the proposition level representing the resulting states after actions have been performed. Valid Plan Found: If the graph plan finds a valid plan, it halts the planning process. Proceeding to the Next Stage: If no valid plan is found, the algorithm determines that the goals are not all achievable in time i and moves to the next stage, OTHER CLASSICAL PLANNING APPROACHES 1. STRIPS (Stanford Research Institute Problem Solver) STRIPS is a planning system developed in the 1970s. It uses a formal language to represent planning problems and solutions. STRIPS is based on the concept of a planning problem as a triple (S, G, A), where: + $ is the initial state + Gis the goal state + A is the set of available actions 2. Plan-Graph-Based Planning Plan-graph-based planning is a planning approach that uses a graph data structure to represent the planning problem. The graph consists of nodes representing states and actions, and edges representing the relationships between them. *plan-graph-based planning is used in planning systems such as Graphplan and Plan-space Planning.3. Plan-Space Planning Plan-space planning is a planning approach that searches the space of Possible Plans to find a solution It uses a planning graph to represent the planning problem and a search algorithm to find a plan *Plan-space planning is used in planning systems such as Plan-space Planning and Planning as Satisfiability (SAT), 4. Planning as Satisfiability (SAT Planning as SAT is a planning approach that reduces the planning problem to a Satisfiability problem and solves it using a SAT solver. It uses a formal language to Tepresent the planning problem and a SAT solver to find a solution. *Planning as SAT is used in planning systems such as Planning as SAT and Planning as Model Checking. 5. Planning as Model Checking Planning as model checking is a planning approach that reduces the planning problem to a model checking problem and solves it using a model checker. It uses a formal language to Tepresent the planning problem and a model checker to find a solution. *Planning as model checking is used in planning systems such as Planning as Model Checking and Planning as Synthesis. 6. HTN (Hierarchical Task Network) Planning HTN planning is a planning approach that uses a hierarchical task network to represent the planning problem.It uses a top-down approach to decompose the planning problem into smaller sub- problems and solve them recursively. *HTN planning is used in planning systems such as HTN Planning and Hierarchical Planning. 7. Planning with Petri Nets Planning with Petri nets is a planning approach that uses Petri nets to represent the planning problem. Petri nets are a formal language for modeling concurrent systems and can be used to represent planning problems. Planning with Petri nets is used in planning systems such as Planning with Petri Nets and Petri Net Planning. ANALYSIS OF PLANNING APPROACHES The analysis of planning approaches in Al involves evaluating and comparing different planning algorithms and techniques to determine their strengths, weaknesses, and suitability for specific planning problems. Types of Analysis: 1. Complexity Analysis: Analyzes the computational complexity of planning algorithms, including time and space complexity. 2. Correctness Analysis: Verifies that the planning algorithm produces correct plans, i.e., plans that achieve the goals and satisfy the constraints. 3. Completeness Analysis: Evaluates the ability of the planning algorithm to find a plan, if one exists. 4. Optimality Analysis: Analyzes the quality of the plans produced by the planning algorithm, including optimality criteria such as plan length, cost, or resource usage. 5. Scalability Analysis: Evaluates the ability of the planning algorithm to handle large planning problems, including the number of actions, states, and goals.Evaluation Metrics: 1. Plan Quality: Measures the quality of the plans produced by the planning algorithm, including optimality criteria such as plan length, cost, or resource usage. 2. Planning Time: Measures the time taken by the planning algorithm to produce a plan. 3. Memory Usage: Measures the memory usage of the planning algorithm. 4. Number of Plans: Measures the number of plans produced by the planning algorithm. 5. Plan Diversity: Measures the diversity of the plans produced by the planning algorithm.