The Optimal Schemes For Transportation Network With Carpooling Based On Speed Limits And Traffic Restriction

The rapid development in modern society brings large scale of vehicle ownerships and travel demands,which results in nonnegligible and increasingly serious traffic congestion.In some cities the congestion has given rise to tremendous loss of economics and time.For alleviating the congestion,various effective travel modes have been preferable.Carpooling,as a travel mode which can reduce the traffic volume without restraining travel demands,is supposed to be advocated.However,the reason why carpooling has not been popular enough at present is the lack of encouragement and infrastructions helpful for carpooling as well as travelers' rejection of carpooling due to their considerations of some factors such as travel costs and privacy.Thus searching for more efficient schemes to encourage carpooling and optimize traffic system is indispensable.Although there existes several schemes such as HOT lanes,road pricing and web-based carpooling platform to encourage carpooling,the effects are still controversial.Furthermore,some policies are difficult to implement due to the responses of the public,for instance,pricing policy.In terms of the development of traffic management policies,various schemes are applied to all around the world.Triggered by the current policies,taking into account the situations with/without HOV lanes we explore three schemes on the basis of traffic restriction and road speed limit for optimization of the traffic system with carpooling to remit congestion and increase social welfare.The main contents include the following sections:1.Introduction of the research background and the significance.This paper reviews the literatures of travel mode split,traffic assignment,carpooling,traffic restriction and road speed limit,demonstrates the state-of-art of the relevant researches,proposes the contents of this paper and describes the research framework.2.Introduction of the definition of traffic management,specific measures and mechanism of the action.In addition,this section illustrates the travel cost functions in the condition of carpooling,traffic restriction,road speed limit respectively,and further states the traffic system by mathematical models describing the equilibriums and mechanism of the action from the perspective of supply and demand.This chapter is the basement of the following sections.3.In a traffic system in the absence of HOV lanes,this chapter introduces the scheme allowing the unrestricted solo driving travelers and carpoolers to enter the restricted area.A variational inequity describes the trip productions,travel mode and route choice.After that the variational inequality is proved to be equivalent to Logit mode split and UE equilibrium,and be solved by previous method such as the block Gauss-Seidel decomposition approach together with the method of diagonalization and successive average algorithm.The effects of various schemes with different restriction areas and proportions are detected in a simple network.The results show that high share of carpooling may not give rise to high social welfare due to the differences among topology of networks and demands.A bi-level programming aiming at optimal restriction area and restriction proportion is designed and applied to a Sioux Falls network with genetic algorithm,and the optimized areas with the constrains in real circumstances are obtained as well as the proportions,which displays the feasibility of this scheme in real world.The results show that this scheme is able to increase social welfare and utilized in real world.4.This chapter detects the scheme allowing some solo driving travelers to enter HOV lanes.In the hypothesis that the network is in peak hours,the travel demand is assumed to be fixed.Firstly the parallel lanes(one link)are considered.A system optimal model and the first-order optimal condition are proposed,showing the marginal utilities of solo driving travelers and carpoolers.Secondly this scheme is extended to a general network.A variational inequality describing the Logit travel mode split and UE equilibrium is applied to a general network with fixed HOV lanes to receive the performance of the network.The results show that this scheme could reduce total travel cost and thereby optimize the traffic system.At the same time,the results verify the smoothing effect of HOV lanes.Then a mathematical programming is provided aiming at minimal the total travel cost,the deployments of HOV lanes.This model is solved by surrogate assisted genetic evolution algorithm and applied to a Sioux Falls network,which proves the effectiveness of the model and algorithm and feasibility of this scheme.5.In the presence of HOV lanes,this chapter introduces the scheme that imposing speed limits on LOV lanes to encourage carpooling and optimize traffic system.Due to the assumption of non-peak hours with fluctuant demands,the travel demands are assumed to be elastic.Firstly a system optimal model with parallel lanes(solo link)is established and the first-order optimal condition is derived.Under this circumstance this section provides the framework to obtain the range of speed limit value.A variational inequility is proposed to describe the Logit mode split and UE equilibrium.Taking a simple general network as an example,the results show that the social welfare and the number of accidents are supposed to be considered simultaneously.A multi-objective model is designed for the optimal scheme and is solved by NSGA II algorithm.This scheme is applied to a simple network and to capture the non-dominant solutions and the bound line of non-dominant solutions.From the conclusions we know that normally the positive impact of speed limit to safety is more apparent than that to social welfare since speed limit decreases road capacity to some extent.Thus road managers must decide the emphasis when enact management policy.Taking a Sioux Falls network as an example displays the feasibility of this scheme in reality.