Bratko, Chapter 11: Basic Problem-Solving Strategies

11.1 Introductory Concepts and Examples
- state space
- start node
- goal condition, satisfied by goal nodes
  - goal(Situation)
- solve(Start,Solution)
  - solve([[c,a,b],[],[]],Solution)
- examples
  - blocks world
  - eight puzzle
  

11.2 Depth-First Search and Iterative Deepening
- depth-first search:
  - solve(State,[State]) :-
      goal(State).
    solve(State,[State|Solution]) :-
      successor(State,State1),
      solve(State1,Solution).
  - eight queens
    - successor(Queens,[Queen|Queens]) :-
        member(Queen,[1,2,3,4,5,6,7,8]),
        noattack(Queen,Queens).
      goal([_,_,_,_,_,_,_,_]).
  - cycle detection [Fig11.7, depthfirst]
  - depth limit [Fig11.8, depthfirst2]
- iterative deepening
  - call depthfirst2 with ever increasing depth limit
  - or generate all paths in increasing order by length [sec11.2]
    - this is actually breadth-first search

11.3 Breadth-First Search
- maintain a list of paths, not a single path
- remove the first path and put successors on the end of the list
- check if first node of first path is a goal node
- use bagof to collect all successors of a node
- breadth-first search with cycle detection [Fig11.10]
  - breadthfirst([[StartNode]],Solution)
  - use of concatentation (conc) is inefficient
- using list differences [Fig11.11]
- using 'path' from iterative deepening

HW
- various searches for blocks world [Exer11.3, p268]