Long-term exposure to high noise level is a major cause of the occupational hearing loss found in industrial workers. By appropriately rotating workers among different workstations, the workers’ daily noise exposures can be reduced. The minimax work assignment problem is intended to find the work assignments for several workers that will minimize the maximum noise exposure that any of the workers will receive. This problem can be viewed as a variation of the minimum-makespan multiprocessor scheduling problem (Pm|-|Cmax) which can be solved by a well-known approximation algorithm called the “longest processing time first (LPT).” In this paper, the modified LPT (M-LPT) swap heuristic is proposed for solving large minimax work assignment problems. After testing on 30 hypothetical problems, the M-LPT swap heuristic is found to be nearly as efficient as the existing genetic algorithm but with much shorter computational time.
DOI 인용 스타일