Quantum annealing task mapping for heterogeneous computing systems

Heterogeneous computing (HC) systems are essential parts of modern-day computing architectures such as cloud, cluster, grid, and edge computing. Many algorithms exist within the classical environment for mapping computational tasks to the HC system's nodes, but this problem is not well explored in the quantum area. In this work, the practicality, accuracy, and computation time of quantum mapping algorithms are compared against eleven classical mapping algorithms. The classical algorithms used for comparison include A-star (A*), Genetic Algorithm (GA), Simulated Annealing (SA), Genetic Simulated Annealing (GSA), Opportunistic Load Balancing (OLB), Minimum Completion Time (MCT), Minimum Execution Time (MET), Tabu, Min-min, Max-min, and Duplex. These algorithms are benchmarked using several different test cases to account for varying system parameters and task characteristics. This study reveals that a quantum mapping algorithm is feasible and can produce results similar to classical algorithms.