A HYBRID LOAD BALANCING APPROACH IN HETEROGENEOUS DISTRIBUTED COMPUTING SYSTEM USING ARTIFICIAL BEE COLONY OPTIMIZATION AND ΒHILL CLIMBING

Abstract

With the advancement in technology and the increasing number of Internet users, it has become necessary to manage servers to prevent overloading or underloading. This requires balancing the load in the network to efficiently utilize resources. In this study, a novel load balancing approach is proposed that augments Artificial Bee Colony (ABC) optimization with a mutation operator and a β-Hill Climbing local improver for task-to-node assignment. The method integrates peer-guided neighbours to produce good partial allocations, employs diversity-preserving micro mutations to avoid premature convergence, and uses a bottleneck-aware β-hill climber that repeatedly moves work from the current max-load node to under-loaded nodes, selecting the best improving micro move at each step. The fitness function reduces makespan, thereby improving resource utilization, throughput, and fairness index. When compared with other existing metaheuristic methods ELBABCEβHC, HDWOA-LBM, and WWO-ACO, the proposed method achieves an improved average resource utilization of 93%, with a reduced average makespan, higher throughput, and fair task allocation to nodes. Across heterogeneous workloads, the algorithm consistently enhances results while scaling favorably as the number of nodes or tasks increases.