Answers are grounded in this paper and cite their sources.
Task scheduling in heterogeneous fog–cloud systems is an NP-hard multi-objective problem requiring simultaneous minimization of makespan, energy consumption, and processing cost, objectives that inherently conflict as workloads scale. Existing metaheuristics often converge prematurely due to poor initialization and fixed exploration–exploitation dynamics. This paper proposes an Enhanced Lemurs Optimizer (ELO) that seeds the exploration–exploitation refinement of the original Lemurs Optimizer (LO) with a problem-aware heuristic schedule under a weighted-sum scalarization framework. Simulations with 100–400 tasks in a heterogeneous fog–cloud environment, using 30 paired runs on shared instances and an equal evaluation budget for all metaheuristics, show that ELO achieves the highest weighted fitness for 200–400 tasks, improving on the original LO by 15.9–19.6% and reducing makespan and energy consumption by up to 24.2% and 21.8%, respectively; for 100 tasks, simulated annealing is marginally but consistently better (by 0.5% in mean fitness). Paired Wilcoxon signed-rank tests confirm that ELO significantly outperforms LO, GA, P2C, and Random at every task size and simulated annealing for 200–400 tasks, and ELO also outperforms NSGA-II when a single schedule is selected with the same weighted-sum rule; an ablation study identifies the heuristic initialization as the main source of the improvement. These results establish ELO as an effective weighted-sum scheduling framework for heterogeneous fog–cloud architectures, with an advantage that grows with problem size.
No takes yet. Share an insight, caveat, or question.
Ababneh et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: