유전알고리듬을 사용한 비정수 지연시간을 갖는 통합 생산/분배 모형

Abstract

Many researchers have developed and evaluated the integrated production/distribution model with time lags. Most of them assume that the time lags are integer multiples of unit time grid. However, in industry practices, integer time lags are very rare and the model developed using the integer time lags is therefore approximating the real system. The differences due to the approximation frequently result in significant accuracy degradation. In this paper, we consider an integrated production/distribution system in supply chain which has non-integer time lags. We focus on a capacitated production planning and capacity allocation problem for the system. We develop a mixed binary integer linear programming model and propose an efficient heuristic procedure using an adaptive genetic algorithm, which utilizes the reduced costs from the LP relaxation of the original model. A regeneration procedure is also developed to evaluate infeasible chromosomes. In order to verify the proposed adaptive genetic algorithm, we tested in terms of the solution accuracy and searching speed during numerical experiments. From the results of experiments, we found that our algorithm can generate the optimal solution within a reasonable computational time.