Heuristic search algorithms for state-space search problems

Cilj ovog rada je izvršiti usporedbu odabranih heurističkih algoritama. U prvom dijelu teoretski su obrađeni pojmovi vezani za pretragu prostora stanja. Opisani su neki primjeri vezani za problem pretrage prostora stanja. U daljnjem radu opisani su neki od osnovnih algoritama pretraživanja. Teoretski je obrađen pojam metaheuristike te su detaljnije opisani algoritmi koji će se implementirati i testirati. Odabrani algoritmi su optimizacija kolonijom mrava, genetski algoritmi te pohlepno slučajno prilagodljivo pretraživanje. Implementacija je napravljena u programskom okruženju NetLogo. Za usporedbu algoritama odabrana je njihova primjena na problemu trgovačkog putnika. Nakon implementacije algoritama izvršeno je njihovo testiranje, dan je pregled rezultata te konačno usporedba algoritama.The goal of this paper is to make a comparison of selected heuristic algorithms. The first part theoretically treats basic concepts of state space search and there have been described some examples related to the problem of state space search. Second part deals with some of the basic search algorithms. The term of metaheuristic has been described and also algorithms that will be implemented and tested. Selected algorithms are ant colony optimization, genetic algorithm and greedy randomized adaptive search. Implementation has been made in the programming environment NetLogo. For better comparison of algorithms they were used on travelling salesman problem. After implementation, algorithms were tested, results were described and finally comparison of algorithm has been made