Development of A Parallel Task Scheduling Solver that Combines A Branch-and-bound Method And Deep Learning

Since the task scheduling problem belongs to the strong NP-hard combinatorial optimization problem, the search time for the optimum solution becomes enormous due to the increase in the scale of the problem. Deep Learning can be applied to this difficult problem. Deep Learning has the advantage that the required time to find a solution is short once learning is completed, but it has the disadvantage that the optimum solution is not always found. Therefore, in this paper, we prototype and evaluate a method for speeding up to find the optimal solution by scheduling that combines the search method based on branch-and-bound method and deep learning