Multi-objective Simulated Annealing for a Maintenance Workforce Scheduling Problem: A case Study

A multi-objective simulated annealing (MOSA) algorithm is described in this chapter to solve a real maintenance workforce scheduling problem (MWSP) aimed at simultaneously minimizing the workforce cost and maximizing the equipment availability. Heavy industry maintenance facilities at aircraft service centres, railroad yards and steel companies must contend with scheduling Preventive Maintenance (PM) tasks to ensure critical equipment remains available (Quan et al., 2007). PM tasks are labour intensive and the workforce that performs those tasks are often highly-paid and highly skilled with different proficiencies, which means the PM tasks scheduling should minimize the workforce costs. Therein lies a dilemma: a small labour force would help control costs, but a small labour force cannot perform many PM tasks per hour—and equipment that is not available does not generate revenue. A long completion time is not cost effective but neither is having too many workforce costs. A proper balance would minimize labour costs while simultaneously finishing all PM tasks in a timely manner. In other words, a trade-off must be made between the workforce costs and a timely completion of all PM tasks. Hence, in most real PM tasks scheduling problems, we encounter the multi-objective optimization. There are very few previous papers focusing on the maintenance workforce scheduling problem. Higgins (1998) formulated the railway track maintenance crew problem as a mathematical program, and then used tabu search algorithms to solve the problem. Ahire et al., (2000) examined the utility of the evolution strategies to solve a MWSP with the aim of minimizing Makespan considering multiple-skills labour and workforce availability constraints. Yanga et al., (2003) formulated an airline maintenance manpower planning problem under a one week planning cycle considering various flexible strategies such as short-term or temporary contracts, trainee, part-time and subcontracted workers. They considered workforces with different types of skills that are grouped into a number of socalled “squads” with different numbers of members (or size). The objective was to minimize the total required manpower while satisfying the demand for every time slot. Quan et al., (2007) used the evolutionary algorithms to solve a multi-objective PM task scheduling problem with the aim of simultaneously minimizing workforce costs and Makespan. Workforce costs consist of the hiring cost of workers required to complete all PM tasks on time as well as the idle time cost. Makespan refers to the total amount of time it takes to O pe n A cc es s D at ab as e w w w .ite ch on lin e. co m