The Optimum Design Of Engineering Structure Based On Gene Expression Programming

Structural optimization design which combined the optimization algorithm withcomputer technology is an important research direction in the modern engineeringdesign field. It reduces the weight and cost of engineering structure and improves thestrength, stiffness, stability, and other performances of structures using structuraloptimization. It is difficult to optimize for large-scale engineering structure because ofthe high nonlinearity and complication using traditional methods.With the development of computer technology and bionics method, ComputingIntelligence provides a new way to structure optimization design in the last decades.Gene Expression Programming (GEP), which appeared in the early21st century, is anew member of genetic calculation family. It has received more and more attentionsbecause of its excellent performance in function finding, times series prediction andother aspects. In this thesis, GEP was introduced and applied to structure optimizationdesign. The details of thesis are following：(1) The updated GEP algorithm is proposed based on the theory of standard GEPand parameter optimization. It improves the efficiencey of standard GEP and preventstrapping into local optimum by introducing the backtracking mechanism. The proposedalgorithm was verified by two examples of parameter optimization.(2) GEP-based structure optimization method is proposed to solve contiuous anddiscrete structure optimization design by combining GEP (GEP-PO, GEP-HZero) withfinite element method for cross setion optimization of truss.(3) To verify the proposed GEP-based structure optimizaiton method, a space trusswith25bars was optimized and analyze the parameters of algorithm. Then, it is appliedto the other two truss examples with10and72bars and comparied with other methods.The results show the proposed method is feasible.(4) All examples are analyzed using ANSYS. The numerical results show theproposed method can be used to structure optimization.