Abstract:

The emergence of human-machine collaboration has adapted to the requirements of big data for high performance computing and complex artificial reasoning, which uses the huge Internet user group and cluster to deal with the increasingly complicated data altogether. In this paper, a bipartite graph matching strategy is proposed to solve the problem of how the crowd and the cluster can collaborate effectively to complete the large data task. The Hopcroft-Karp algorithm of bipartite graph matching not only enhances and extends the Hungarian algorithm, but also considers the field of adaptive segmentation tasks, the degree of association, and the evaluation of the background and ability of the crowd to maximize the matching between the crowd and the segmented task group. The algorithm calculates each influence factor after each match and optimizes the next match, making the best match between the crowd and the task. Through the experiment, the accuracy of the task completion is verified to be the highest.