Basic AI search algorithms implemented in R.

• Estado.R defines a general representation of a state in a AI search problem;
• buscaDesinformada.R implements Uninformed Search algorithms: breadth-first, depth-first and uniform cost)
• buscaInformada.R implements Informed Search algorithms: greedy and A*.

Script in Canibais.R implements Estado.R for the problem of the Missionaries and Canibals, defining the generation of new states given a set of operators and the heuristica evaluation for this particular problem.

Script exemploCanibais.R instantiates a version of the Missionaries and Canibals (3 missionaries, 3 canibals and 1 boat, all starting on the left margin of the river) and runs all the available searches.

Defining initial and objective nodes:

initial <- Canibais(desc = c(M = 3, C = 3, B = 1))

objective <- Canibais()
objective$desc <- c(M = 0, C = 0, B = 0)

Executing a breadth-first search is as simple as calling the buscaEmLargura(initial, objective) function, which returns the path obtained from the initial node to the objective node as a list object:

buscaEmLargura(initial, objective)

The other available searches work similarly.

For the informed search, best-first is the executed algorithm. Besides the initial and objective nodes, the function expects a parameter indicating the approach to be used in the search: "Greedy" must be passed for the greedy search and "AEstrela" for the A* search. If no value is passed, the function uses the defined default of "AEstrela".

buscaBestFirst(inicial, objetivo, "Greedy")

Obs: It is worth noting that the problem of 3 Missionaries and 3 Canibals has a small and simple search tree, so the path returned for all serach algorithms is the same.