Research On Cross Business Task Sharing And Allocation Method In Urban Computing

Urban computing is a process to solve the challenges faced about cities(such as environmental degradation and traffic congestion)by continuously acquiring,integrating and analyzing a variety of heterogeneous big data in cities.It combines ubiquitous sensing technology,efficient data management and analysis algorithms,and novel visualization technology to improve people’s quality of life,protect the environment and promote urban operation efficiency.Urban computing includes many services,such as urban travel,urban logistics,urban energy consumption and urban finance,which provide great convenience for users’ life.The smart media connecting urban computing and users is the rapidly developing Internet platform,such as Meituan,Di Di and Ctrip.These platforms not only provide services for users,but also bring more benefits to online workers on the Internet(such as online car Hailing drivers and parcel couriers).Most existing Internet platforms contain many different services.For example,Meituan not only provides users with ordering business and taxi business,but also provides users with medical,fitness and other services.However,at present,most Internet platforms focus on the distribution of single businesses.There is no cross integration between multiple businesses,and each business needs to be arranged to a class of professional workers on the platform.This distribution model has two defects: one is that it will cause a serious waste of human and space resources,and the other is that it will aggravate traffic congestion and environmental pollution.Therefore,this paper proposes a novel cross business task sharing and allocation problem in urban computing.Its core problem is how to allocate tasks fairly and reasonably for workers on the Internet platform.The optimization goal is to maximize the total income of the Internet platform on the premise of ensuring a certain degree of fairness.Aiming at the problem of cross business task sharing and allocation in urban computing,this paper takes two common tasks of urban travel and urban logistics as examples to carry out a series of research,and its main contributions are as follows:(1)Cross business Task Matching(CTM)is defined,and a two-stage solution framework is proposed.The first stage is package site planning.Firstly,a weight model that can weigh the traffic flow and the distance between stations is constructed to represent the advantages and disadvantages of the paths between stations,and then a path planning algorithm is proposed to plan an optimal delivery path for each package.The second stage is cross-business task sharing and assignment.Firstly,a cross-business task packaging strategy is proposed,which is based on the location and time of passengers and packages,and then uses a game theory-based fairness perception algorithm to allocate task packages and workers.(2)The two-stage solution framework is optimized.For the package site planning stage,an Adaptive Station Planning(ASP)algorithm based on deep reinforcement learning is proposed.The algorithm adaptively plans an optimal site path for each package by capturing the changes of traffic flow in the real-time environment,so as to improve the success rate of package delivery and increase the total income of the platform.Aiming at the stage of cross business task sharing and allocation,a Cooperative aware Game Theoretical Matching(CGTM)algorithm is proposed.The algorithm integrates the packaging process into the best response framework,so that the Nash equilibrium state finally achieved by the game theory not only ensures the income equity of workers,but also makes multiple cross business tasks assigned to workers have task packages with high cooperation quality scores.Extensive experimental results show that ASP algorithm can effectively improve the total revenue and task completion rate of the platform,while CGTM algorithm has excellent performance in improving matching efficiency.(3)Aiming at the cross business task scheduling process in urban computing,a Cross business Scheduling and Monitoring Platform(CSMP)is designed and developed.The platform can effectively monitor the scheduling process and make real-time statistics and summary of urban travel business and urban logistics business.At the same time,it provides three friendly user interfaces for display to help users(urban computing decision-makers or data analysts)conduct real-time monitoring and analysis on the cross business task allocation process and quality in urban computing.