001 package aima.search.framework;
002
003 import java.util.ArrayList;
004 import java.util.List;
005 import aima.basic.Agent;
006 import aima.basic.Percept;
007 import aima.util.Util;
008
009 /**
010 * Artificial Intelligence A Modern Approach (2nd Edition): Figure 3.1, page 61.
011 * <code>
012 * function SIMPLE-PROBLEM-SOLVING-AGENT(percept) returns an action
013 * inputs: percept, a percept
014 * static: seq, an action sequence, initially empty
015 * state, some description of the current world state
016 * goal, a goal, initially null
017 * problem, a problem formulation
018 *
019 * state <- UPDATE-STATE(state, percept)
020 * if seq is empty then do
021 * goal <- FORMULATE-GOAL(state)
022 * problem <- FORMULATE-PROBLEM(state, goal)
023 * seq <- SEARCH(problem)
024 * action <- FIRST(seq)
025 * seq <- REST(seq)
026 * </code>
027 * Figure 3.1 A simple problem-solving agent. It first formulates a goal and a problem,
028 * searches for a sequence of actions that would solve the problem, and then executes the actions
029 * one at a time. When this is complete, it formulates another goal and starts over. Note that
030 * when it is executing the sequence it ignores its percepts: it assumes that the solution it has
031 * found will always work.
032 */
033
034 /**
035 * @author Ciaran O'Reilly
036 *
037 */
038 public abstract class SimpleProblemSolvingAgent extends Agent {
039
040 // seq, an action sequence, initially empty
041 private List<String> seq = new ArrayList<String>();
042
043 //
044 private boolean formulateGoalsIndefinitely = true;
045
046 private int maxGoalsToFormulate = 1;
047
048 private int goalsFormulated = 0;
049
050 public SimpleProblemSolvingAgent() {
051 formulateGoalsIndefinitely = true;
052 }
053
054 public SimpleProblemSolvingAgent(int maxGoalsToFormulate) {
055 formulateGoalsIndefinitely = false;
056 this.maxGoalsToFormulate = maxGoalsToFormulate;
057 }
058
059 // function SIMPLE-PROBLEM-SOLVING-AGENT(percept) returns an action
060 @Override
061 public String execute(Percept p) {
062 String action = NO_OP;
063
064 // state <- UPDATE-STATE(state, percept)
065 Object state = updateState(p);
066 // if seq is empty then do
067 if (0 == seq.size()) {
068 if (formulateGoalsIndefinitely
069 || goalsFormulated < maxGoalsToFormulate) {
070 if (goalsFormulated > 0) {
071 notifyViewOfMetrics();
072 }
073 // goal <- FORMULATE-GOAL(state)
074 Object goal = formulateGoal();
075 goalsFormulated++;
076 // problem <- FORMULATE-PROBLEM(state, goal)
077 Problem problem = formulateProblem(goal);
078 // seq <- SEARCH(problem)
079 seq.addAll(search(problem));
080 if (0 == seq.size()) {
081 // Unable to identify a path
082 seq.add(NO_OP);
083 }
084 } else {
085 // Agent no longer wishes to
086 // achieve any more goals
087 die();
088 notifyViewOfMetrics();
089 }
090 }
091
092 if (seq.size() > 0) {
093 // action <- FIRST(seq)
094 action = Util.first(seq);
095 // seq <- REST(seq)
096 seq = Util.rest(seq);
097 }
098
099 return action;
100 }
101
102 //
103 // PROTECTED METHODS
104 //
105 protected abstract Object updateState(Percept p);
106
107 protected abstract Object formulateGoal();
108
109 protected abstract Problem formulateProblem(Object goal);
110
111 protected abstract List<String> search(Problem problem);
112
113 protected abstract void notifyViewOfMetrics();
114 }