Dynamic Task Allocation Of Personal Rapid Transit Based On Knowledge Graph

With the improvement of people’s living standards,vehicles in the city are also increasing year by year.Although many measures have been taken,the traffic congestion problem has not been improved fundamentally.At present,the problems faced by ground transportation come from the sharing of road rights,and the subway,light rail and other transportation modes with special tracks have greatly limited the service area due to factors such as construction cost,operation cost and occupied space.In this context,it is particularly important to build an efficient and low-carbon intelligent transportation system.Therefore,some scholars have proposed Personal Rapid Transit.PRT system is a kind of intelligent transportation system,which integrates the advantages of existing public transportation.It can achieve fast and direct “point to point” and play the role of “capillary” to realize the connection between the first kilometer and the last kilometer.Therefore,it is of great significance to study the PRT system to solve the problem of traffic congestion.With the continuous improvement of the key technologies and hardware equipment of the PRT system,the PRT system will play an important role in the public transport system.Once it is put into operation in a large scale,a large number of dynamic passenger flow will put forward higher requirements for the collaborative task allocation of the PRT system.Therefore,this paper studies the collaborative task allocation problem of PRT system,establishes the model of alliance formation and task allocation,and uses the improved particle swarm algorithm to solve it.The main research contents are as follows :Firstly,the collaborative control idea of PRT system based on multi-agent network is proposed,the multi-agent network model of PRT system is established,and the knowledge representation of vehicle task allocation problem based on knowledge graph is constructed.Secondly,for the NP-Hard problem of collaborative task allocation in PRT system,considering the changes of passenger flow at each station and the distribution of vehicle resources,the arrival and queuing models of passenger flow are established based on queuing theory,and the blackboard-based coalition communication model is proposed to solve the key problems in the formation of coalition.Thirdly,the multi-objective task allocation model of PRT system is established aiming at the utilization rate of vehicle mileage and passenger waiting time,and the improved EBPSO algorithm is used to solve the model.Without losing the mileage utilization rate,compared with the standard particle swarm algorithm,the average waiting time and the longest average waiting time are reduced by 47.95% and 41.31%,respectively.Compared with the particle swarm algorithm that only improves the fitness function,the average waiting time and the longest average waiting time are reduced by 11.17% and 14.85%,respectively.Finally,taking the surrounding area of the school as an example,this paper designs the road network layout model of the PRT public transport system in the region and conducts experiments,and analyzes the indicators of system efficiency and user service level in the regional public transport system based on demand response.Experiments show that the algorithm in this paper is feasible and effective,which can quickly and effectively solve the collaborative task allocation problem of PRT system,make the system realize intelligent operation,improve task execution efficiency and system efficiency,and provide reference scheme for the actual operation of PRT system.