Chaotic Genetic Algorithm (CGA) And Structural Optimization

The research of the theory and algorithm of optimization may trace back to theforties of the twentieth century, which is flourished owing to the development ofcomputer. Many conventional numeric optimization algorithms based on alternate arebrought forward, which are extensively applied in the fields of industry and economy.However, along with its further development and kinds of problems arising in thepractical application, the research of the theory and algorithm of optimization is stillurgently necessary.Genetic algorithm (GA) is a bionics algorithm, which simulates the biologictheory and is recently widely concerned in the fields of computer science andoptimization. As an algorithm of global optimization, GA has many remarkablecharacteristics such as simplicity, common applicability, strong robustness, beingapplicable in parallel processing and vast application area. Although GA has formedits basic structure as a searching algorithm and displayed its advantages by solvingdifferent problems, it also reveals many theoretic and technical shortcomings. It'seffective to combine GA with local searching method in order to improve the GA'sperformances. Chaos is a universal phenomenon in nonlinear systems, with intensesensitivity to starting values. Hence, searching with chaotic variables is easier todepart from local solution.Structural optimization means searching the best solution under constraintconditions with optimal methods that are divided into indirect methods and directmethods. It includes three phases: establishing mathematical model, selecting areasonable and effective computational method, and composing computer program.This paper researches and analyses GA's developing actuality, basic structure,main characteristics, improving techniques and application in structural optimization,and structural optimization's researching background, developing actuality andcharacteristics. A kind of chaotic generic algorithm (CGA) based on real coding andchaotic theory is proposed and tested on the common test functions in this paper.Optimal model and basic process of GA applied in structural optimization aredisplayed. Also the optimization of 25-bar truss and 72-bar truss proves the algorithmis effective.