160 likes | 301 Views
Introduction to Artificial Intelligence CS 438 Spring 2008. Today AIMA, Ch. 3 Problem Solving Agents State space search Programming Assignment Thursday AIMA, Ch. 3 Uniformed Search. Can your IA make you money as a gold farmer?. Goal-based Problem Solving.
E N D
Introduction to Artificial IntelligenceCS 438 Spring 2008 • Today • AIMA, Ch. 3 • Problem Solving Agents • State space search • Programming Assignment • Thursday • AIMA, Ch. 3 • Uniformed Search Can your IA make you money as a gold farmer?
Goal-based Problem Solving • To develop an IA there two major concerns • KR • Search Method • Assume for right now that knowledge is encoded in some form that it can be easily retrieved and applied
Goal Formulation • Goal • A state of the environment that meets some desirable property or properties • Examples • Chess: checkmate (opponents king cannot avoid capture) • Path finding: being in a specific geographic location • Robot Vacuum: Clear floor • Goals may include factors that determine which solutions are more desirable than others • Speed, shortest path, safety
Goal Formulation • Before you can decide what to do you must determine what it is you are trying to do • Take an “intentional stance” D. Dennett • “Goals help organize the behavior by limiting the objectives that the agent is trying to achieve” p. 60 • Given all possible actions to take, some can be rejected outright because they are not relevant of the agent reaching its goals.
State-space Search • Search • The activity of looking for a sequence of actions that solves (achieves) the goal (goal state) • State-space • Defined by the initial state, the actions the agent can take to go from one state to the other, and goal state
State-space search • Path • Any sequence of action that leads from one state to another • Solution • A path starting at the initial state and leading to the goal state • Path cost • Sum of the cost of each action • g(n) cost of path from initial state to state n • Note that path cost differs from “search cost”, which refers to the computational complexity of the search algorithm
Problem Formation • Initial State • State the agent starts in • Actions available to the agent • Defines actions that allow IA to transform one state into another • Successor function: S(x) given state x returns set of new states given each applicable action (action-state pairs) • Goal test • Determines if a state meets the specific properties of the goal • Path cost • Function assigns a cost to a solution path
Single-state problem formulation • initial state: "at Arad" • actions or successor functionS(x) = set of action–state pairs • S(Arad) = {<Arad Zerind, Zerind>, … } • goal test • x = "at Bucharest" • path cost (additive) • sum of distances, number of actions executed, etc.
Example: The 8-puzzle • states? • actions? • goal test? • path cost?
Example: The 8-puzzle • states?locations of tiles • actions?move blank left, right, up, down • goal test?= goal state (given) • path cost? 1 per move
Example: robotic assembly • states?: real-valued coordinates of robot joint angles parts of the object to be assembled • actions?: continuous motions of robot joints • goal test?: complete assembly • path cost?: time to execute
Example: Water Jug Problem • Goal formulation: measure precisely 2 gallons of water • Problem formulation • Two jugs • 4 gallon jug with x amount of water • 3 gallon jug with y amount of water • Initial state • (0,0) : both jugs empty • Goal • (2, y) or (x, 2) • Path cost: 1 unit for each pouring action
Water Jug Problem • Actions:
Search • Expanding a state • Generating new states by applying possible (valid) actions to current state using the successor function S(x) • Search Tree • Root is the initial state • Each expanded state is a search node • Search Node • Encodes the state, parent node, action applied, depth, and path cost