Two Metaheuristics for Multiobjective Stochastic Combinatorial Optimization

摘要

Two general-purpose metaheuristic algorithms for solving multiobjective stochastic combinatorial optimization problems are introduced: SP-ACO (based on the Ant Colony Optimization paradigm) which combines the previously developed algorithms S-ACO and P-ACO, and SPSA, which extends Pareto Simulated Annealing to the stochastic case. Both approaches are tested on random instances of a TSP with time windows and stochastic service times. ant colony optimization; combinatorial optimization; multi-objective decision analysis; simulated annealing; stochastic optimization