Scheduling and Optimization of Dynamic Complex Discrete Events Based on Ant Colony Optimization

An ant colony optimization (ACO) based on Multi-agent approach for project scheduling and optimization is presented. Several new features that improve the converge speed and precision of ACO in general are proposed and evaluated: (1) new updating strategy of pheromone; (2) importing trust mechanism as heuristic information; (3) multi-node searching strategy; (4) problem solving by multi-agent alliance negotiation. Based on these improvements, we presented architecture, steps and applications of ACO based on multi-agent. We tested this algorithm on a real product-scheduling problem. Compared to basic ACO for this problem solving, the algorithm presented in this paper performed best on average.