Asynchronous parallel hyper-heuristic method for large-scale optimization problem

The invention discloses an asynchronous parallel hyper-heuristic method aiming at a large-scale optimization problem. The method comprises the following steps of performing initialization operation; sub-nodes evolve the population in parallel; asynchronous self-adaptive communication between the main node and the sub-nodes is carried out; the main node updates the population distribution table, the algorithm probability table and the population according to the comparison result; solving a large-scale task optimization problem; according to the method, the convergence degree of the population in different evolution stages is considered, the hierarchical model is established, and different algorithms can be fairly compared and evaluated in the same evolution stage, so that the underlying algorithm is effectively scheduled; the multi-strategy characteristic of a hyper-heuristic method and the time advantage of parallel computing are organically combined through a hierarchical model, and the performance of the a