An Application of Genetic Algorithms to Production Scheduling Considering the Division of Orders

In recent years, genetic algorithms (GA) which simulate the process of natural evolution have gained attention as a means of solving optimization problems, beginning with scheduling problems. This paper intends for the problem of division of orders (unit quantity of manufacturing units comprising one type of model) in production process scheduling. In actual production activity, since some orders cannot be completed by the delivery date due to the production situation at the factory or the customer's requirements, scheduling is conducted to divide an order among many machines. The introduction of order division also has considerable effects on improving the operation rate of production equipment and reducing total production time. However, there are many combinations for determining which orders should be divided and how many machines they should be divided among, and no effective solution has yet been reported. We suppose that GA is effective for solving the combination problem mentioned above. This paper proposes an individual description of GA which realizes optimum order division, and reports experimental results of our simulation.