Ant Colony System algorithm solving a Thermal Generator Maintenance Scheduling Problem

Abstract

The maintenance scheduling problem of thermal generators is a large-scale combinatorial optimization with constraints. In this paper an Ant Colony System (ACS) algorithm, one of the Ant Colony Optimization (ACO) algorithms, is proposed for the maintenance scheduling problem. This ant colony optimization method allows the “agents” of an ant colony to deposit a small amount of pheromone trail to every path that has been explored, thus passing on to the other agents the information concerning the best solution. With the iterations we construct the final solution. This method is called “positive feedback”. The basic optimization routine is reinforced with the introduction of elitist ants who make the best solution stronger. The algorithm is applied to a real-scale system, and further experimenting leads to results that are commented.