Application Of Hybrid Algorithms Based On Ant Colony Optimization And Artificial Fish Swarm For Batch Processing Machine With Non-Identical Job Sizes

In the real life and production,whether it is machine processing,parts manufacturing,or cargo shipments,aerospace,need to solve the scheduling problem.Scheduling problem is not only a combination of optimization problems,but also has a widely application background,it has played an important role to improve the overall social labor productivity,resource use efficiency and reduce production costs,has amount of research.The batch scheduling problem is an important extension to the classical scheduling problem.It is a new type of dispatch problem originating from the semiconductor production process.Batch scheduling has very important theoretical and economic value.In this paper,the problem of batch scheduling is NP-hard problem.Simple and efficient algorithm is the main direction of batch scheduling research.The main algorithm used in this paper is ant colony algorithm and fish swarm algorithm.After a brief introduction of ant colony algorithm and fish swarm algorithm,according to the characteristics of the algorithm and the limitation of vision,an improved fish swarm algorithm is proposed:Through the dynamic change of the field of view,the algorithm improves the early search width and the later convergence rate,and further improves the efficiency of the algorithm.Through the experiment and result analysis,the improved fish swarm algorithm is more efficient than the traditional fish swarm algorithm.Based on batch scheduling problem characteristics,combined with the advantages and disadvantages between ant colony algorithm and fish algorithm,two hybrid algorithms are proposed:through the combination of the crowding degree factor,prevented the ant colony algorithm from falling into the local extremum at the early stage,which leads to the algorithm early maturing.The algorithm has the global optimization ability and can find a better value.The main problem in this paper is to solve the problem of single machine batch scheduling,the workpiece is non-identical sizes,and only one processing machine.The algorithm parameters need to be reset for the specific problem,the pheromone definition,heuristic information and pheromone initialization are improved,and the fish algorithm is also adjusted accordingly.In order to ensure the persuasiveness and effectiveness of the experiment,the number of experiments,the size of the workpiece processing,and the length of the workpiece processing time are classified.In order to intuitively and comprehensively compare the experimental results,we applied the concept of batch utilization and load rate.The utilization of the batch reflects the degree of utilization of the machine capacity in the batch processing time;the batch load rate reflects the degree of waste in the overall processing time.According to the results of experiment,the performance of the ant colony algorithm is better than the fish swarm algorithm in the process of algorithm optimization,but the early mature of the ant colony algorithm leads to the local optimum as the optimal result.However,combining the ant colony algorithm with the fish swarm algorithm,combining the crowding factor with the ant colony algorithm,it can effectively avoid early mature and has help for reducing the searching time in finding the optimal solution.Through the comparison of the first hybrid algorithm and the second hybrid algorithm,the second hybrid algorithm has a better efficiency for the problem that both the number of workpieces and iterations is small.The first hybrid algorithm has a higher performance for the larger number of parts and iterations times.