A Topology-aware Multi-objective Hiking Optimization Algorithm (MOHOA) for Sustainable Task Scheduling in Heterogeneous Cloud Environments

The core of sustainable cloud computing is to have efficient task scheduling, and a fine balance between performance and resource consumption is necessary.The presented work introduces Topology-Aware Multi-Objective Hiking Optimization Algorithm (MOHOA), a novel metaheuristic used for the optimization of the task scheduling in the heterogeneous cloud environments.Unlike the conventional methods, MOHOA integrates Tobler Hiking Function (THF) to dynamically adjust the search velocity based on the slope of the fitness landscape and is effective at both crossing complex search spaces and local optima avoidance.The algorithm employs Pareto-based Archive (PArchive) with crowding distance mechanism in handling the conflicting objectives: Makespan, Energy Consumption and Total Execution Cost.The performance of MOHA has been carefully compared to state of the art algorithms, like MOEA/D, NSGA-II, and customized metaheuristics, in diverse settings with as many as 1000 tasks.The statistical validation based on 30 independent runs and also based on multi-objective indicators (Hypervolume, Spread, and IGD) validates that MOHOA is significantly better than its competitors.In terms of large-scale workloads, MOHOA particularly made a reduces makespan (16.62%) and also enhanced power efficiency; the experimental results demonstrate that MOHOA achieves competitive performance across different task sizes, particularly in large-scale scenarios.At the 1000-task level, MOHOA reduces power consumption by approximately 11.09% compared to MMWOA, and by about 9.13% compared to MCS, confirming its effectiveness in improving energy-aware scheduling.The algorithm is a good competitor of sustainable cloud infrastructure management because sensitivity research validates its capacity and scalability to sustain a large variety of excellent trade-off solutions.