AI: Logic in AI 2
2 likes2,971 views
A simple planning agent uses percepts from the environment to build a model of the current state and calls a planning algorithm to generate a plan to achieve its goal. Practical planning involves restricting the language used to define problems, using specialized planners rather than general theorem provers, and adopting hierarchical decomposition to store and retrieve abstract plans from a library. A solution is a fully instantiated, totally ordered plan that guarantees achieving the goal.
AI: Logic in AI 2
- 1. Logic in AI 2
- 2. A simple Planning AgentA simple planning agent uses the percepts provided by the environment to build a complete and correct model of the current world state, after which , to achieve its goal, it calls a suitable planning algorithm (which we will call IDEAL-PLANNER) to generate a plan of action.
- 3. Problem solving and Planningby a simple planning agentBasic elements of a search-based problem-solver are Representation of actions, Representation of states, Represents of goals and Representation of plans.
- 4. Components of practical planning(1) Restrict the language with which PLANNER we define problems. With a restrictivelanguage, there are fewer possible solutions to search through.(2) Use a special-purpose algorithm called a planner rather than a general-purposetheorem prover to search for a solution.
- 5. Basic representation for planningLEAST COMMITMENT this principle says that one should only make choices about things that you currently working. PARTIAL ORDER A planner that can represent plans in some steps are ordered with respect to each other and other steps are unordered is called a partial order planner.LINEARIZATION A totally ordered plan that is derived from a plan P by adding ordering constraints is called a linearization of P.
- 6. What is a PLAN?A plan is formally defined as a data structure consisting of the following four components:Set of plan stepsSet of step ordering constraintsSet of variable binding constraintsSet of casual links
- 7. What is a Solution?A solution is a plan that an agent can execute, and that guarantees achievement of the goal. If we wanted to make it really easy to check that a plan is a solution, we could insist that only fully instantiated, totally ordered plans can be solutions.
- 8. How to Resolve threats in planning?Resolve now with an equality constraintResolve now with an inequality constraintResolve Later
- 9. Knowledge Engineering for planningDecide what to talk about.Decide on a vocabulary of conditions (literals), operators, and objects.Encode operators for the domain.Encode a description of the specific problem instance.Pose problems to the planner and get back plans.
- 10. Practical PlanningHierarchical decompositionThe practical planners have adopted the idea of hierarchical decomposition: that an abstract operator can be decomposed into a group of steps that forms a plan that implements the operator. These decompositions can be stored in a library of plans and retrieved as needed
- 11. Analysis of Hierarchical DecompositionAbstract solution A plan that contains abstract operators, but is consistent and complete once an abstract solution is found we can prune away all other abstract plans from the search tree. This property is the downward solution property.We can prune away all the descendants of any inconsistent abstract plan. This is called theupward solution property
- 12. Resource constraints in planningUsing measures in planningThe solution is to introduce numeric-valued measures. Measures such as the price of gas are realities with which the planner must deal, but over which it j has little control. Other measures, such as Cash and Gas Level, are treated as resources that can be produced and consumed.Temporal constraintsIn most ways, time can be treated like any other resource. The initial state specifies a start time for the plan.
- 13. Visit more self help tutorialsPick a tutorial of your choice and browse through it at your own pace.The tutorials section is free, self-guiding and will not involve any additional support.Visit us at www.dataminingtools.net