Research On Comprehensive Integrated Optimization Modeling For Urban Electric Transit Networks

Improvement in the urban electric transit system is able to effectively alleviate numerous social problems such as traffic congestion and environmental pollution.Nevertheless,there are still some issues,such as irrational layout of transit network,low service frequency and low punctuality of public transport services.Moreover,the limited driving range and the shortage and high cost of charging infrastructure are the main barriers which restrict the development of electric buses.In view of the above objectives that need to be improved,this study presents comprehensive integrated optimization models for urban electric transit network design problem(ETNDP)according to different modeling perspectives to simultaneously optimize the transit network layout,the service frequency and the charging station location.The corresponding heuristic algorithms are developed to solve the optimization models and tested on the transit network in an urban region of a city in China.Afterward,a comparative analysis of the effect of various policies on improving the operation efficiency of the transit network and the overall travel efficiency of all passengers is made by the proposed methodology.The contents of this dissertation are as follows:Firstly,the use of electric buses and the construction of charging stations will add the operating cost,thus increasing the financial burden of government.This study proposes a transit network operating cost optimization model with the objective of minimizing total operating cost of the transit network for one day to simultaneously optimize the transit network layout,the service frequency and the charging station location.An artificial fish swarm algorithm(AFSA)with the crossover and mutation operators is developed to solve the proposed optimization model to overcome the slow convergence rate in the later phase of iterations and to ensure the diversity of swarm.It is found that the presented optimization model and AFSA are able to appropriately optimize the urban electric transit network with effective control of operating cost,while guaranteeing adequate operating buses to meet all passenger demands and satisfying the recharging demands of all operating buses.Moreover,when solving the ETNDP,the conflicting interests of passengers and operators need to be considered.This study presents a transit network operating cost and passenger travel efficiency bi-objective optimization model by simultaneously optimizing the layout of transit network,the service frequency and the location of charging stations with the objectives of minimizing total operating cost of the transit network for one day and total generalized cost for all passengers.In consideration of the weighted sum method in which the setting of weights hugely depends on the subjectivity and empiricism,this research develops a Pareto Artificial Fish Swarm Algorithm(PAFSA)to solve the proposed bi-objective optimization model to coordinate the conflicts between the passengers and operators.It is demonstrated that the proposed bi-objective optimization model and PAFSA are able to rationally optimize the urban electric transit network from the perspective of coordinating different stakeholders to reflect the decision-making effect under the principle of different interests.Furthermore,as the number of chargers installed in each charging station has a great impact on the optimization of ETNDP,this study proposes a transit network operating cost and passenger travel efficiency integrated optimization model with the aim of simultaneously designing the layout of transit network,the service frequency,the location of charging stations and the number of chargers to be installed in each charging station.A PAFSA embedded with the genetic algorithm(GA)is presented to solve the proposed integrated optimization model.The PAFSA is used as the main framework of the solution algorithm to design the layouts of bus routes and the associated service frequencies by minimizing total operating cost of the transit network for one day and minimizing total generalized cost for all passengers.During which,the effect of real-time transit information(RTI)on the selections of travel path and departure time is considered to achieve the more realistic description of the travel behaviors of passengers.The GA is taken as a subroutine of PAFSA to decide the locations and sizes of charging stations which depend on the weighted sum of total construction and operational costs of all charging stations and total deadheading distance.The dynamic system simulation method is used to calculate the fitness function of GA.It is found that a significant decreases of total waiting time and total transfer penalty for all passengers with the use of RTI.Meanwhile,the total construction cost and the total operational cost of all charging stations are both reduced because of using GA.That is,the proposed model solved by the newly developed algorithm is able to optimize urban electric transit network more rationally from the perspective of system optimization.Lastly,sensitivity analyses are conducted to understand the impacts of average operating speed,transfer penalty,RTI query ratio,battery capacity,average electricity price and average bus ticket on the optimization results of the proposed integrated optimization model.According to the sensitivity analyses,seven optimization strategies are proposed from the views of network planning,route operation,vehicle technology and financial subsidies.The compared results show that accelerating the development of battery technology will most effectively improve the operation efficiency of the transit network in the study area.In contrast,expanding the service coverage of real-time transit information in the study area will significantly improve the satisfactory level of all passengers.Moreover,the application of all optimization strategies is able to maximize the operation efficiency of transit network and the overall travel efficiency of all passengers but it could also be economically the most expensive and difficult to carry out due to the different policy implementers.There are 36 figures,8 tables and 171 references in the dissertation.