A heuristic genetic algorithm for solving complex safety-based work assignment problems

This paper presents a heuristic genetic algorithm (GA) for finding the work assignments for a group of workers in which the maximum daily noise exposure level that any of the workers receives is minimized. Since the classical partially matched crossover (PMX) and insertion mutation operators require extensive computation time to solve large work assignment problems, new heuristic crossover and mutation operators are proposed. These heuristic operators are intended not only to reduce the computation time but also to reduce the number of generations that GA requires to converge to the optimal solution. In this paper, the scope of the work assignment problem has been expanded to cover both the balanced (the numbers of workers and workstations are equal) and unbalanced (the number of workers is greater than that of workstations) problems. Additionally, the feasibility of assigning a worker to a workstation (worker–workstation restriction) or so-called assignment restriction is considered. Eight work assignment problems are analyzed. Three problems are balanced, unrestricted problems, three are unbalanced, unrestricted problems, one is balanced, restricted problem, and one is unbalanced, restricted problem. Each problem is solved using both the conventional GA (with the PMX, crossover and the insertion mutation) and the heuristic GA. The results show that the heuristic GA can yield the optimal or near-optimal solution much faster than the conventional GA and it also requires lesser number of generations.