Task Planning Of Crowdsourcing Logistics In Dynamic Environment

Crowdsourcing refers to outsourcing specific tasks to non-employee.With the development of Information and Communication Technology(ICT)and mobile Internet,a new spatial crowdsourcing which involving location based in crowdsourcing is emerging.Spatial crowdsourcing plays an important role in many domains,such as intelligent transportation system(ITS),public safety,and environmental monitoring.Crowdsourcing logistics is an innovative spatial crowdsourcing application enabling the public to join urban logistics distribution,which makes full use of social idle resources and reduces the logistics cost.Crowdsourcing logistics planning problem is to plan the most suitable distribution tasks for the deliverymen according to the spatial distributional characteristics of delivery tasks and deliverymen.Different from traditional Vehicle Routing Problem(VRP),crowdsourcing logistics platform is online where the distribution of delivery and deliveryman would change dynamically in real-time,thus the route planning is a great challenge.Here,two dynamic task planning modes are considered: Worker Selection Mode(WST)and Server Assignment Mode(SAT).The main study consists of three parts as follows.(1)In terms of WST,assuming the deliveryman select the delivery tasks independently,the first crowdsourcing logistics task planning model and solving method are presented.A dynamic optimization framework is presented to maximize the number of distribution tasks for a deliveryman.Four greedy algorithms(including EDR,NNR,NSR,and HDR)and the Tabu heuristics are presented to solve the problem.The experiments show that the solution quality of NSR algorithm is higher and relatively stable in different scenarios.The competition ratio of the Tabu heuristics can reach about 0.9,after iterative optimization.In this mode,the task planning results are recommended as auxiliary information to help deliverymen accomplish the choice.(2)In terms of SAT,assuming the crowdsourcing logistics system assign delivery tasks to deliveryman,another crowdsourcing logistics task planning model is presented.The objective is to minimize the travel cost and delivery delay of all deliverymen.A hybrid metaheuristic method using adaptive large neighborhood search and Tabu is designed.The hybrid metaheuristic method integrates a distribution task prediction algorithm and a dispatcher scheduling algorithm is presented to improve the quality of task planning.The experiments show that the performance of the presented hybrid metaheuristic method is at least 4.53% higher than the Tabu.The scheduling algorithm can effectively reduce the delivery delay.(3)A prototype system for crowdsourcing logistics planning is designed.In detail,the system requirement analysis and module design are carried out.The prototype system of crowdsourcing logistics is implemented by integrating the presented dynamic task planning algorithms based on the Web framework,which further verified the feasibility of the proposed models and algorithms.