Shape Scheme and Size Discrete Optimum Design of Plane Steel Trusses Using Improved Genetic Algorithm

The objective of this study is the development of a scheme and discrete optimum design algorithm, which is based on the genetic algorithm. The algorithm can perform both scheme and size optimum designs of plane trusses. The developed Scheme genetic algorithm was implemented in a computer program. For the optimum design, the objective function is the weight of structures and the constraints are limits on loads and serviceability. The basic search method for the optimum design is the genetic algorithm. The algorithm is known to be very efficient for the discrete optimization. However, its application to the complicated structures has been limited because of the extreme time need for a number of structural analyses. This study solves the problem by introducing the size & scheme genetic algorithm operators into the genetic algorithm. The genetic process virtually takes no time. However, the evolutionary process requires a tremendous amount of time for a number of structural analyses. Therefore, the application of the genetic algorithm to the complicated structures is extremely difficult, if not impossible. The scheme genetic algorithm operators was introduced to overcome the problem and to complement the evolutionary process. It is very efficient in the approximate analyses and scheme and size optimization of plane trusses structures and considerably reduces structural analysis time. Scheme and size discrete optimum combined into the genetic algorithm is what makes the practical discrete optimum design of plane fusses structures possible. The efficiency and validity of the developed discrete optimum design algorithm was verified by applying the algorithm to various optimum design examples: plane pratt, howe and warren truss.