Multiobjective ant colony search algorithm for optimal electrical distribution system strategical planning

In this paper, a dynamic MultiObjective, MO, algorithm based on the Ant Colony Search, the MultiObjective Ant colony Search Algorithm, MOACS, is presented. The application domain is that of dynamic planning for electrical distribution systems. A time horizon of H years has been considered during which the distribution system will be modified according to the new internal (loads) and external (market, reliability, power quality..) requirements. In this scenario, the objectives the Authors consider most important for utilities in strategical planning are: the quality requirement connected to the decrease of the expected number of interruptions per year and customer, in the considered time frame, and the choice for the lowest cost strategy. The Authors have formulated on purpose a new dynamic optimization algorithm to treat hard MO problems such as the one of strategical planning. The algorithm is a MO version of the Ant Colony Search based on the concept of Pareto optimality. Namely, it works with as many colonies of ants as many objectives the problem presents and it is conceptually divided in two different phases a forward phase, in which single objectives and local increments are prized and a backward phase in which Pareto optimal solutions are prized.