Research indexJournal article
Local Elite-Guided Collaborative Optimization Algorithm for Continuous Distributed Constraint Optimization Problems
Shi, Yu, Chen
Journal of Computing and Electronic Information Management27 Apr 2026Open access
Local Elite-Guided Collaborative Optimization Algorithm for Continuous Distributed Constraint Optimization Problems
Shi, Yu, Chen · Journal of Computing and Electronic Information Management · 2026
Abstract
To address the issues of premature convergence to local optima and slow convergence speed in Continuous Distributed Constraint Optimization Problems (C-DCOPs), this paper proposes a Local Elite-Guided Collaborative Optimization (EGA-LC) algorithm. Utilizing a Breadth-First Search (BFS) pseudo-tree, the algorithm constructs a distributed population structure partitioned into ten sub-populations for parallel search, each further classified into dominant, non-dominant, and local dominant sub-populations based on fitness. Subsequently, a differential evolution strategy is designed to guide individual updates through the synergistic influence of global elites and local dominant solutions, enabling deep mining of high-quality local constraint information. Furthermore, a periodic cross-population recombination and reassignment mechanism is introduced, which leverages historical global optimal solutions to accelerate convergence and prevents evolution stagnation by reallocating sub-populations. Experimental results on four benchmark problem sets demonstrate that EGA-LC significantly outperforms five state-of-the-art algorithms in terms of both solution quality and convergence speed.