Batch Scheduling Algorithm For Parallel Machines Based On Multi-variety, Small Batch And Order

The production model of multi-variety, small batch and make to orders is very important in the manufacture industry. Flexible Manufacturing Cell is used in production model of multi-variety and middle-small batch.FMC is the manufacturing system with high flexibility and high automation. In recent years, FMC is applied in the multi-variety and middle-small batch's factory.Due time is the lifeline of the make to order enterprise. It's difficult to take into account the requirement of the different order delivery when making the workshop plan. In the production environment of FMC, in order to reduce the costs of unit and improve equipment utilization, different orders can be composed of several parts which contain the same type,and this parts can be processed successively. The program of batching and scheduling directly affects the economic efficiency and reputation of enterprises.A large tire mold manufacturing enterprises as the background, An integrated problem was studied for lot-sizing strategy and parallel machines scheduling, in which the constraints of order's due time, arrival time and setup time were considered in this paper. At first, in the environment of order's arrival time is determined, to reduce the manufacturing costs and tardiness penalty costs, a single model was built up to describe the whole problem. A genetic-simulated annealing algorithm with heuristic rules(GASA) was proposed. The heuristic rule was introduced for shorten the length of chromosome and increasing the search speed. Lot-sizing strategy was derived globally by genetic algorithm. Then the optimal scheduler was searched by simulated annealing algorithm in present lot-sizing strategy. Secondly, to examine the effective of GASA, some real production examples of this mold manufacturing enterprise were used. The result shows that the algorithm is effective convergence in the acceptable time. Under the same computing time, computational simulations and comparisons based on five kinds of application cases with different sizes were provided. Results demonstrated that the proposed algorithm obtained better results on the larger size problems. It is shown out that the scheduling model and the algorithm are more reliable and effective. Thirdly, the mathematical model for batch scheduling was built up, and taking into account the arrival time of orders in production is uncertainly in real production environment. The research shows that the process of orders'arrival consistent with a Poisson process. In the environment of the probability density function of random variables was curtained, expectations model was established. So the model can deal with as a deterministic model. This problem can be solved by GASA which mentioned. Through large-scale numerical simulation experiments, the result shows that it's effective to solve the problem of orders'arrival time uncertainly by GASA. Problem solving process is not sensitive to the orders'arrival time. At last, the simulation model was built up by advanced modeling and simulation tools named eM-Plant with the method of Orient Object. In order to restore the real production environment, every object and action controlled by SimTalk in eM-Plant. It is shown out that the GASA algorithm are more reliable and effective in the reality through the result of running the simulation model and computer by the numerical simulation.