AI based single and hybrid meta heuristic algorithms for energy and resource aware scheduling in cloud fog computing environment

Abstract

Scheduling in cloud and fog environments is a complex and NP-hard problem, where newlineallocating tasks across a distributed, heterogeneous network of resources becomes computationally newlineintensive as system scale and workload dynamics increase. Despite advancements newlinein metaheuristic algorithms designed to approximate near-optimal scheduling newlinesolutions, the unpredictability of cloud-fog networks presents significant challenges. newlineResources in fog environments are often constrained and distributed, serving to latencysensitive newlineapplications such as IoT and real-time analytics. These settings demand adaptive, newlineenergy-efficient scheduling that metaheuristics are inadequate to provide, as they newlinecannot guarantee optimal solutions within polynomial time. Consequently, latency, energy newlineusage, and bandwidth constraints compound the scheduling problem, resulting in newlinepotential degradation in performance and reduced overall efficiency. newlineWe have proposed various single and hybrid Meta-Heuristic (MH) algorithms for newlineenergy-aware and resource-aware scheduling in a cloud-fog environment to overcome newlinethese scheduling challenges. Our first approach, an Energy Aware Electric Fish Optimization newlineAlgorithm (EAEFA), uses Electric Fish Optimization (EFO) to balance dynamic newlineworkload demands by meeting energy-aware objectives, including makespan, newlineenergy consumption, total cost, and throughput. Experimental results show that EAEFA newlinereduces makespan time by 15.13%, energy consumption by 15.45%, and total cost by newline9.65% while improving throughput by 12.25%. Although EAEFA is effective for local newlineoptima, achieving the global optimal solution remains challenging. To address this, we newlinepropose a hybrid energy-aware algorithm, the Electric Earthworm Optimization Algorithm newline(EEOA), which combines EFO for local optimization with the Earthworm Optimization newlineAlgorithm (EOA) for global optimization. EEOA further improves makespan newlineby 14.78%, energy consumption by 12.35%, and total cost by 10.53%, while enhancing newlinethroughput by 11.54%. newlineFor resource-aware scheduling, we introduce th