Research On Resource Scheduling Algorithm And Its Applications

In order to complete complex combat missions successfully, the method which organizesmultiple combat units with relative cooperation mechanisms is very necessary. For a multiplecombat units collaborative system, it can’t work the way it’s supposed to if there isn’t a suitablecooperation mechanisms with it. This may even leads to a bad result because of the unsuitablestrategy which assigned too much resources to an unimportant target while there is few resourcecan be assigned to other targets.First of all, we analyzed the crucial factors of resources assignment problem for multiple unitsfight in coordination from the perspective of cooperative system. Based on these analyses and theactual needs of combat missions, as well as the ability of combat units, we made a reasonablemodel-Resource Scheduling Model of Multi-units Co-operation Fight(RSM-MCF) for theproblem which could describe right constrains of the problem instead of a simple model which istoo simple for such a complex problem. Our model is a Multi-Objective OptimizationProblem(MOOP) essentially. It considered both the quality and the cost of a combat mission.Compared to other model such as MILP and TSP, RSM-MCF has more restrains that is related tothe problem, also is more objective, leads to more practical significance.Secondly, in order to find the solution more efficiently, we did some useful try to the researchof multi-objective optimization algorithm. Because of Differential Evolution(DE) algorithm’sadvantage for solving continuous optimization problem, we chose DE to solve our model. In orderto avoid the situation of falling into local optimum when solve the MOOP with DE, we designed abidirectional search mechanism which can improve the ability of local search of the DE. We alsodesigned a Multi-population mechanism for DE, which can reduce the risk of local optimum, aswell as makes the Pareto fronts more evenly distributed.At last, we designed a simulation platform system for cooperate combat of multiple units. Thesystem does reasonable functional partition and module design, and it regard resource assignmentalgorithm as core. By display the whole battleground dynamically, the system verifies our modeland algorithm’s feasibility.