Multi-Objective Monarch Butterfly Optimization algorithm for agri-food workflow scheduling in fog–cloud

The explosion of data and connected objects has encouraged the development of distributed computing environments such as cloud, fog and edge, facilitating real-time data processing. In the agri-food sector, this evolution is reinforced by the emergence of the Industrial Internet of Things (IIoT) and fog–cloud computing. One of the major challenges of these environments is the efficient scheduling of applications while respecting the SLA principle. This article deals with the multi-objective optimization of agri-food workflow scheduling using an improved version of the Monarch Butterfly Optimization (MO-MBO) algorithm. Our approach integrates Pareto dominance, a greedy strategy, and a self-adaptive strategy to effectively handle four conflicting objectives: makespan, cost, energy, and latency. Simulations carried out on FogWorkflowSim for the agri-food health monitoring workflow demonstrate that MO-MBO is energy-efficient (68.63% compared to the genetic algorithm (GA), 67.78% compared to the particle swarm optimization algorithm (PSO), and 85.48% compared to the Non Dominated Sorting Genetic Algorithm-II (NSGA-II)) and improves latency (0.35% for GA, 0.14% for PSO, and 0.10% for NSGA-II). However, cost increases marginally, by 0.85% for GA, 0.29% for PSO, and 0.10% for NSGA-II. Makespan, meanwhile, increases by 31.41% versus GA, 29.27% versus PSO, and 10.67% compared versus NSGA-II. The results highlight the ability of MO-MBO to effectively balance multiple, conflicting objectives, making it a promising solution for energy-efficient, low-latency scheduling in fog–cloud systems.