Sizing Optimization of Truss Structures using a Hybridized Genetic Algorithm

TL;DR

This paper introduces a hybrid genetic algorithm tailored for optimizing truss structures, combining global exploration with local exploitation to efficiently handle complex, nonlinear design problems with many variables.

Contribution

A novel hybrid algorithm that enhances genetic algorithms with local search for efficient truss structure optimization, outperforming existing methods.

Findings

Superior performance in weight reduction of trusses

Effective handling of large variable and constraint spaces

Outperforms traditional optimization techniques

Abstract

This paper presents a genetic-based hybrid algorithm that combines the exploration power of Genetic Algorithm (GA) with the exploitation capacity of a phenotypical probabilistic local search algorithm. Though not limited to a certain class of optimization problems, the proposed algorithm has been "fine-tuned" to work particularly efficiently on the optimal design of planar and space structures, a class of problems characterized by the large number of design variables and constraints, high degree of non-linearity and multitude of local minima. The proposed algorithm has been applied to the skeletal weight reduction of various planar and spatial trusses and shown to be superior in all of the cases,