Ant colony optimization for resource-constrained project scheduling

An ant colony optimization approach (ACO) for the resource-constrained project schedul-ing problem (RCPSP) is presented. Combi-nations of two pheromone evaluation meth-ods are used by the ants to nd new solutions. We tested our ACO algorithm on a set of large benchmark problems from the PSPLIB. Compared to several other heuristics for the RCPSP including genetic algorithms, simu-lated annealing, tabu search, and dierent sampling methods our algorithm performed best on the average. For some test instances the algorithm was able to nd new best solu-tions.