Research On Dynamic Vehicle Routing Problems Considering Parking

In recent years,with the rapid growth of car ownership,Self-Driving Trips and Travels(SDTT)therefore show a simultaneous growth trend.The ever-intensifying conflict between supply and demands of parking caused by the lag-behind infrastructure construction makes time-consuming parking procedure one of the most serious challenges of SDTT.Navigation and parking decision are two independent topics in existing literature.The broadly used vehicle-mounted navigators are only capable of static path planning,though convenient to SDTT to some extent,it demonstrates serious limitations in Chinese metropolises where parking lots are usually insufficient.Taking the characteristics of dynamic relationships between parking lots determination and path planning into consideration,this thesis comprehensively analyzed parking lots determination procedure and its related influences and carried out the research about automatic dynamic navigation based on the model of path planning integrated parking lots determination.Major works and results are as follows:(1)The representative research achievements on path planning and parking optimization have been summarized.Analyzed were the influencing factors of parking determination and the quantitative modeling method about the factors.As for the drawbacks of the disassociation of navigation and parking lots determination,and the poor performance of dynamic path planning,a targeted solution and a related technical path were proposed.(2)An improved genetic algorithm was proposed to solve the problems of"premature convergence"and poor local search ability in traditional genetic algorithm.These improvements included;?a population initialization strategy combining random selection and terminal tendency concept;? a cross point location method based on spatial distance proximity;?a local search strategy based node fitness.By taking the road network data of Fuzhou as samples,the superiority of the proposed genetic algorithm in accuracy,efficiency and stability was verified.(3)A multi-objective synthetic decision-making method with K-dominance constraint was proposed.This method optimized parking lots determination procedure by considering"quality"and"quantity"of target value of parking decision influential factors.Contrast experiments were implemented with traditional multi-objective decision method and multiple Logit optimization methods based on utility theory to verify the predominance of the proposed method in accuracy and robustness.(4)An overall path planning model considering parking was proposed.Focusing on the coherence between parking lots determination and path planning in driving behavior,their different interaction in fore-proceeding and proceeding were considered to construct two overall planning models respectively;meanwhile,a from-whole-to-local strategy was applied to address the inefficiency of the proceeding time calculation.Empirical research cases in Fuzhou urban area are used to verify the efficiency and robustness of the proposed model.