Research on Computing Task Scheduling Under Comprehensive Computing Power Measurement

Cloud-network convergence has become a paradigm for new networks, and computing force networks have emerged as a result, aiming to cope with the increasing demand for computing power. However, due to the weak coupling between task scheduling and computing measurement in computing clusters, as well as the coarse granularity of computing measurement, it is difficult to efficiently allocate tasks to the most appropriate computing nodes based on accurate and reliable computing resource evaluation. Therefore, in this paper, we propose an optimized ant colony algorithm T-ACO (topsis ant colony algorithm) that incorporates comprehensive computing measurements, with the goal of shortening the task execution time while minimizing the task cost. Firstly, the static resource status of computing nodes is scored using the comprehensive evaluation method TOPSIS to derive the computing, storage, and cost scores of computing nodes, which are incorporated into the initialization pheromone and pheromone update of the ACO algorithm, and then the real-time resource status of the nodes is added to optimize the heuristic information function. The simulation experiments show that the T-ACO algorithm has good accuracy and real-time performance, which improves the performance compared to the two comparison algorithms.