Innovative parallel grasshopper optimization algorithm for reliability optimization

TL;DR

A new optimization algorithm called p-GOA improves reliability optimization by using parallel strategies for global and local search.

Contribution

The p-GOA introduces a parallel cooperative strategy combining migration and mutation for better reliability optimization.

Findings

p-GOA outperforms existing methods in finding reliable systems and converging faster on four reliability problems.

The parallel approach balances global exploration and local refinement without penalty functions.

The algorithm effectively handles real-world engineering constraints like cost and resource limits.

Abstract

This study introduces a novel Parallel Grasshopper Optimization Algorithm (p-GOA), specifically designed to address reliability optimization problems. Although several hybrid algorithms exist in this field, the proposed p-GOA distinctly differs through its parallel cooperative strategy. Unlike sequential methods that apply techniques one after another, p-GOA simultaneously divides the population into two groups operating in parallel: one group employs a migration strategy (SOMA) for broad global exploration of the search space, while the other utilizes a mutation operator (NUMO) for focused local refinement of solutions. This dual-strategy parallel operation creates achieving a stronger balance between global exploration and local refinement, while a smart penalty-free method naturally steers the search toward workable solutions. When tested on four well-known reliability problems, the results demonstrate that our method consistently finds more reliable systems and converges faster than existing approaches, demonstrating its effectiveness in handling real-world engineering constraints.

● This study introduces a Parallel Grasshopper Optimization Algorithm (p-GOA) that integrates GOA, SOMA, and a Non-Uniform Mutation Operator (NUMO). It employs mutation, migration, and a parallel approach to efficiently explore both feasible and near-feasible regions without relying on penalty functions.

● The p-GOA dividing the population into two parallel groups—one updated using SOMA-based migration and the other using NUMO-based mutation. This dual-strategy, simultaneous processing not only accelerates convergence but also strengthens the balance between global search and local optimization.

● Specifically targets reliability optimization problems, particularly redundancy allocation issues where components must meet specific reliability and resource consumption (cost, weight, volume) constraints.

Image, graphical abstract

Figures