Optimization of large scale maintenance networks

This paper describes a comparison between a single phase evolutionary programming algorithm for the fixed destination multi-depot multiple travelling salesman problem with multiple tours problem (mmTSP) and the generalized tabu search algorithm. This optimization problem widely appears in the field of logistics mostly in connection with maintenance networks. In these types of systems, there are devices or sometimes whole sites. In this article they called objects and handle as a single node in a graph. These objects require periodic examination, but the problem is that they are scattered over a wide area, often over a country. Another problem is that the examinations cannot be done in an arbitrary period there is a minimum time between the examinations. The first part of the article describes the problem and shows some literature study. In the next part the mathematical model of the system is presented. The following chapter discusses the constraints of the optimization. The constraints are implemented with penalty functions as shown in the next chapter, which followed by the introduction of the general tabu search algorithm. Finally, the comparison of the previously developed evolutionary programming algorithm is presented. The researches presented in this article show the advantages of the developed evolutionary algorithm over the tabu search algorithm. In the future, application of other algorithms like the firefly algorithms is planned in this problem area.