Strategy for the evolutionary optimization of reinforced concrete frames based on parallel populations evolving

Abstract

Abstract An algorithm has been developed for optimization of reinforced concrete frames with the possibility of forming a structure with a subsequent selection of reasonable section sizes and reinforcement for it. A modified genetic algorithm is used as the main search tool. The overall optimization process is divided into two stages. At the first stage, the topology is synthesized by randomly excluding groups of structural members as well as changing the position in space of the remaining groups. At the second stage, local parallel evolving genetic algorithms are performed for the formed topologies. With a common set of variable parameters, the migration of individuals between parallel populations is not allowed. An example of optimization of the multi-story building frame is considered.