Research And Application Of Complex Resource Scheduling Problem Based On Variant Population Evolutionary Algorithm

With the rapid development of national productivity,resource scheduling has become a core issue in many fields.The essence of resource scheduling is to allocate and schedule limited resources reasonably,while complex resource scheduling is to explore solutions suitable for multiple resource scheduling scenarios on the basis of single resource scheduling.The challenge of complex resource scheduling comes from many factors,including the finiteness of resources and the diversity of demands,as well as the differences between different resources,etc.The rationality and advantages and disadvantages of scheduling schemes have an important impact on the cost,efficiency,and quality of resource scheduling.Variant population evolution,as a classical strategy in heuristic algorithms,enables more efficient iteration of operators and is a preferred solution for solving complex resource scheduling problems.The thesis mainly establishes corresponding mathematical models for two types of problems,namely,shared bus scheduling and preventive resource scheduling,and designs a variant population evolution algorithm based on the pyramid model,combined with the underlying scheduling algorithm and applied to these two types of problems.As a product of the sharing economy model,the shared bus is an important part of the smart city.The thesis mainly establishes a shared bus scheduling model with multiple cycles and time windows to minimize the scheduling cost by taking passengers with travel demand in a specified set of locations within the same city as the research object.By designing a sequence flow encoding to represent the vehicle scheduling scheme,a variant population evolution algorithm VPPM-VNS based on the pyramid model with a variable neighborhood search strategy is proposed to guide the population to iterate at a more efficient rate.Through experimental comparison with the other five algorithms,the VPPM-VNS algorithm can obtain more stable and better solution quality.In the process of prevention and control of highly infectious diseases,resource scheduling is a very critical issue.The thesis mainly proposes a multi-objective evolutionary algorithm MEAPM-VNS based on pyramid model with variable neighborhood search for the prevention and control resource allocation problem-oriented to dynamic spatiotemporal contact networks,with the multiple objectives of minimizing prevention and control resource usage,minimizing prevention and control duration and maximizing fairness of prevention and control resource allocation.By an experimental comparison with the other five algorithms,MEAPM-VNS generates an undominated solution set with better solution quality.Finally,a complex resource scheduling platform is developed for the algorithms for the above two problems.The requirements of the functions to be realized by the system are analyzed in detail.The design ideas are expounded from the aspects of architecture design,detailed design,and database design,and the platform’s operation flow is displayed.