The Research on Improved PSO Algorithm-Based Decision-Making over Maintenance Works

Abstract

Maintenance works decision-making is a scientific approach to addressing the conflict between the supply of maintenance resources and the demand for it. Whether for routine or emergency maintenance works, maintenance works decision-making is always beneficial to enhance their efficiency massively. Therefore, in the field of maintenance works decision-making, the key problem lies in how to make the deployment of maintenance inventory and the assignment of maintenance works optimal under the constrains like usage expenses, availability, spare parts fill rate and so on. This paper starts with the multi-target problem and the Particle Swarm optimization algorithm, and then proposes the improved multi-target PSO algorithm. The rationale is that, fussy adjustment is made to the inertia weight and acceleration factor, to increase the number of sub-groups formed by the learning particle swarm. Meanwhile, the particles with an optimal location are generated in the new particles of the subgroups for the next-step calculation of particle location, to compare and update the non-inferior solutions in the external files. As shown by the comparative experiment of test functions, the algorithm proposed in this paper could improve the classic PSO algorithm significantly in terms of the number of solutions and their distribution. Finally, some assumptions are made to model the decision-making over the practical maintenance works, which indicates that this algorithm is quick to work out a high-quality feasible solution. It is effective to support the practice of maintenance works, showing its feasibility and practicality.