Research On Flexible Scheduling Of Public Transport Considering Multiple Vehicle Types And Operation Models

With the rapid development of economy and the improvement of people's living standards,people's travel modes have become more and more diversified,convenient and economical.People's yearning for a better life also derives new requirements for transportation.At present,the transit models in most Chinese cities are not manifold enough to meet the growing needs concerning people's life and work.Demand Responsive Transit(DRT)is a new type of public transport service mode,which usually provides public transport services in low-density population areas or areas with specific needs.At present,with respect to the research on DRT at home and abroad,practical problems are often elucidated by the theoritical analysis embracing vehicle type,road network,vehicle scheduling modaity and on the like.The parallel schedule of multiple operation modes is rarely taken into account,and flexible combination schedule of multiple vehicle types is also rarely considered.In light of the above research background,this thesis proposes a new mode of flexible bus schedule,which involves multiple vehicle types and multiple operation modes.This thesis revisits the literature concerning both DRT and path planning at home and abroad.Next,the key issues of DRT flexible scheduling are illustrated from five aspects: vehicle type selection,key issues of direct scheduling mode,key issues of joint passenger scheduling mode,travel time window,and site clustering method;then,the workflow of DRT flexible scheduling system service is introduced.In terms of modeling,firstly,the application scenario of DRT flexible scheduling model is described and its service scenario is depicted;secondly,the classic DRT direct scheduling model and DRT collaborative combined scheduling model are modeled;finally,the double decision variables of vehicle type and path are innovatively set up so that flexible scheduling model of public transport considering multiple vehicle types and multiple operation modes is constructed.To address this issue,the model is profoundly analyzed,in which the more complicated connection problem is simplified into two-step solution by means of step-by-step solution.At the same time,the same problem is integrated andclassified by which two types of vehicle path problems are summarized,which are MDVRP solution with different starting and ending points of fixed destinations and MDVRP solution with different starting and ending points of multiple destinations.Later,HGA hybrid genetic algorithm(NNS + 2-opt + destination dimension reduction operator + genetic algorithm)and hgaco(hybrid genetic algorithm + ant colony algorithm)are used to solve the model innovatively,and small-scale practical cases are also applied to verify the algorithm.In the aspect of algorithm performance analysis,the comparative analysis of genetic algorithm,hybrid genetic algorithm and hybrid genetic ant colony algorithm is conducted.The optimal solution and convergence speed are compared and analyzed under different data sets,which verifies the performance enhancement of the improved hybrid algorithm.Finally,with that example from the southwest area to the central passenger station in Jiexi city,the feasibility and scientificity of DRT flexible scheduling model are verified in practical application.The results reveal that the study of flexible bus scheduling considering multiple models and operation modes is practical and can further optimize the bus service level and enterprise cost.