Network Optimization Of Crowdsourcing Distribution Based On K-means-Genetic Algorithm

With the rise of "lazy economy", people’s demands for 020 services are growing day by day. There appeared many 020 platforms which provide a variety of 020 service for life. Relative to other requirements, demands of 020 services are huge. Higher request on distribution timeliness and safety brings about the 020 platform using self-built distribution team which can’t fulfil the demands of "the last three kilometers" distribution. So we need to develop a new way of distribution: Crowdsourcing distribution. However, the crowdsourcing distribution personnel are from social capacity resource overwhelmingly. From the perspective of the 020 platform, it can’t realize the standardized management like self-built distribution team. From a view of personal point, the crowdsourcing distribution personnel don’t have professional distribution personnel’s quality and experience. Therefore, the current crowdsourcing distribution of the 020 service delivery exists problems, including distribution delayed and no one to accept the order. We need some relevant optimizations for crowdsourcing distribution.The paper studies on crowdsourcing distribution network, for the purpose of optimizing the crowdsourcing distribution network. First, we analyze the status of crowdsourcing distribution network. Based on the results of questionnaire survey, we analyze process, methods and existing problems of crowdsourcing distribution. The key problems of optimizing distribution route and merging crowdsourcing distribution demands are proposed in the paper to solve problems of distribution delayed and no one to accept the order in the crowdsourcing distribution network. In view of the distribution range of the crowdsourcing distribution network within 3 km, we use the K-means clustering algorithm to partition the distribution area. Considering the need to ensure the timeliness and fairness of crowdsourcing distribution, combined with the crowdsourcing distribution characteristics to determine the number of clusters. The regional distribution division standards prescribe as the distance of various demand notes minimum and the total demand of each cluster equal roughly. After the completion of the distribution region divided, we analyze the single cluster and confirm the crowdsourcing distribution network. In considering of the dynamic characteristics of crowdsourcing distribution demands, we take the single distribution income maximum of crowdsourcing distribution personnel as the objective function, establish crowdsourcing distribution route basic optimization model and crowdsourcing distribution route dynamic optimization model to solve the key problems of optimizing distribution route and merging crowdsourcing distribution demands. And we select genetic algorithm to solve the models.Finally, to D’s case, the crowdsourcing distribution network is analyzed by the case of D crowdsourcing distribution platform. Through empirical analysis, we prove the models have played a very good role in optimizing for the crowdsourcing distribution network.