The Algorithm Research On Scheduling Optimization In AS/RS

In the modern logistics system, the Automated Storage and RetrievalSystem (AS/RS), as the core and hub of logistics system, is the key to achievethe rationalization of logistics system. It is very important in modern logistics.From the current structure of the AS/RS, it’s hardware technology has been verywell, but people still do not pay attention to scheduling optimization in AS/RS,and it is not a fully developed area. Therefore, the use of AS/RS is limited.Under this condition, in order to improve the operating efficiency of the AS/RS,to increase the competitive advantage, the research on scheduling optimizationin AS/RS has very significant application value and practical significance.The efficiency of AS/RS depends mainly on the location assignment andloading/unloading scheduling, because they are core content of scheduling inAS/RS. If the location assignment and loading/unloading scheduling arearranging reasonable, storage space will be fully used and the shelf force will bein a good condition. At the same time，it can reduce the stacker’s workload,short the distance and operating time when loading/unloading goods, reduce thecost of the logistics business, increase income. So the main subject of this paperis studying the problem of location assignment and loading/unloadingscheduling in AS/RS.When doing location assignment in AS/RS, the relevant principles of cargopartition and location assignment are summarized firstly. Mathematical modelof a single row and multiple rows of location assignment are established underboth consider of the stability of the shelf and the efficiency of loading/unloading.Then the Polychromatic Sets (PS) and Particle Swarm Optimization Algorithm(PSO) are introducing, and a solution based on PS and PSO is proposed.According to the theory of PS, the rack area is zoned, the number of which isequal to the category of goods. Hybird PSO is used to assign goods location interms of the category and quantity of input goods in each subarea. Thus, theproblem of location optimization is solved and the validity and feasibility of themodel is examined by an instance. When doing loading/unloading scheduling in AS/RS, at first, the conceptand the principle of loading/unloading scheduling are summarized,mathematical model of loading/unloading scheduling in multiple rows areestablished. When the model is solved, in order to avoid premature convergenceof conventional Genetic Algorithm (GA), an improved discrete particle swarmoptimization algorithm based on PS is presented. During the solution process,polychromatic sets matrices is used to assign loading/unloading goods locationreasonably in the particle position of PSO to increase the primary particlesquality, thus the search performance of the algorithm and optimize results canbe improved. Then in the situation of the iterative number is fewer, comparedwith GA through examples, the effectiveness of the convergence speed andresults are verified.