Multi-step Intelligent Computational Methods For Satellite Module Layout Design

Multi-step intelliget computational methods are proposed for Satellite Module LayoutDesign Problem(SMLDP).According to features of SMLDP, it is fast and effective to solveSMLDP by making full use of heuristic methods,while the intelligent optimizationalgorithms are improved to escape the local optimal and accelerate the convergencerate.Numerical experiments show that such methods are very effective.Firstly, start with the research background of layout problems to lead to relevantbackground and research progress of SMLDP, and a brief analysis of the currentdifficulties of solving SMLDP, and then briefly describe research work done in this articlefor SMLDP.Secondly, solutions to SMLDP mainly contain heuristic methods and co-evolutionaryapproach, while heuristic methods include order-based positioning method and full packingoptimization method. Then multi-step method based on rules related to constraints isintroduced to reduce multi-dimensional constraints.Subsequently, to solve the prematurity of evolutionary algorithm on two-dimensionalSMLDP, the multi-quantum states quantum-inspired evolutionary algorithm (MQSQIEA)proposed in this thesis is beneficial to keeping the population diversity, which combineswith the heuristics based on the order-based positioning technique. Quantum-inspiredevolutionary algorithm is improved by multi-quantum states coding, the convergencecriterion based on the average convergence probability, the observation method based onthe taboo strategy and heuristic information and the dynamic quantum evolutionarystrategy to enhance the efficiency of optimizing the layout sequence, while the positioningprobability function introduced to the positioning rule is employed to improve the solutionquality. Numerical experimental results show that the proposed method can effectivelysolve two-dimensional SMLDP.Finally, For complex multi-constraint coupling SMLDP, cooperative co-evolutionaryartificial bee colony algorithm(CCABC) combines improved artificial bee colony algorithm with the cooperative collaborative framework. The multi-step strategy based onthe coupling relationship of constraints from low to high is to reduce the complexity due tomultiple constraints. Numerical experimental results show that the proposed method caneffectively solve multi-constraint coupling SMLDP.In short, multi-step intelliget computational methods are proposed for2D and3DSMLDP respectively. Numerical experimental results show they are effective to SMLDP.