Research On Scheduling Model Andalgorithms Of Logistics Matching Problem

The rapid development of business activities speeds up the development of logisticsindustry. With the advancement of global economy and electronic business, the proportion ofcosts of logistics is bigger and bigger, thus the effectiveness problem of logistics has becomemore and more prominent. Compared with developed countries, China’s logistics performanceindex is low, which shows that our country’s logistics efficiency still needs to be improved. Thedata shows that, the empty loading ratio of vehicles in highway transportation of our country isabout40%, while it is only10%in the United States. High empty loading ratio not only bringsabout a waste of resources, but also has great influence on traffic, environment etc.With the development of information technology, communication technology and artificialintelligence technique, fully tapping potential of existing logistics resources and improving useefficiency of logistics resources conform to the current development requirements of our country.It is well known that vehicles are the core of logistics resources, thus taking effective measuresto increase the vehicle load factors is of vital importance for improving logistics efficiency.Concepts like collaborative logistics and transport cooperation put forward by scholars both athome and abroad, as well as the joint distribution advice of the government department are allput forward around how to improve the vehicle load factors.In the field of logistics, the problems of vehicle capacity collocation, fully excavatinglogistics capacity can be summed into the class of logistics carpooling matching problem, namedcarpooling matching problem in short. How to choose the right line to make cars exert maximumcapacity in the process of carrying, and minimize average costs at the same time, is the researchpurpose of logistics matching problem. Obviously, expenses incurred are shared to allparticipants. If scheduled properly, for logistics participants, it will undoubtedly greatly reducethe cost of output and improve economic benefit, which is of great significance for solving thehigh transportation cost problem existing in current logistics industry of our country. In the fieldof manned transport, similar problem also exists. The world developed countries like America,Germany, and Singapore etc. began to try carpooling in the1970s and their policy, infrastructureand technology are relatively mature. In recent years, with the rapid development of economy,the vehicle population is growing rapidly thus bringing about a series of social problems, such astraffic jam, environment and noise pollution etc. Civil society and the academia also beginsexploring policy and technical means to solve this problem, concepts of carpooling and freeriding are put forward and many first-tier cities begin to have a try.From the perspective of circulation, cargo and passengers can be combined to be understoodas the service demand, and the purpose of matching is under the premise of meeting or trying tomeet service demand, to transport from the source to the destination with less cost. Nature ofthese two is relatively consistent and can be attributed to carpooling matching problem. To sumup, the logistics matching problem as the key technical problem in the field of circulation hasimportant theoretical research and practical application value, based on which research of this paper unfolds.Main research contents of this paper include the following aspects.1. Research on the definitive single vehicle logistics carpooling matching problem inlogistics matching problem. Based on service demand, vehicle positional relationship, timewindow constraint etc. allocate several service demands to one car and study how vehicles canachieve carpooling at low costs and try to improve carpooling success rate. As for the vehiclescheduling problem in modern logistics, the definitive single vehicle logistics carpoolingmatching problem in logistics scheduling problem----SVLRMP problem is put forward. Itsdefinition is formalized and objective function and constraint conditions are determined; thematching degree clustering heuristic algorithm based on prior clustering----MDCA algorithm isproposed; in specific path optimization process, the competition-prey co-evolutionary hybridmodel among multiple species is built, based on the uniformity index among communitiesguiding the execution of co-evolution genetic algorithm, and applying it to single vehiclelogistics carpooling matching problem. Numerical experiment shows that the MDCA algorithmcan choose service requirement with high accuracy, get carpooling solution in short time, andeffectively reduce the costs that vehicles bear.2. On the basis of single vehicle logistics carpooling matching, consider cooperationamong vehicles and extend single vehicle problem to definitive multi-vehicles. Studymulti-vehicles logistics carpooling matching problem; study proper clustering scheme of servicedemands based on matching situation of vehicles with same service requirements and allocateservice demands solely to one car; study multi-vehicles logistics carpooling matchingoptimization plan. Aiming at vehicle co-scheduling problem in modern logistics, on the basis ofSVLRMP problem, definitive multi-vehicle logistics carpooling matching problem in logisticsscheduling---MVLRMP problem is put forward. Its definition is formalized and objectivefunction and constraint conditions are determined; aiming at the requirement of multi-vehicleco-scheduling, the two-stage clustering heuristic algorithm----TSCA algorithm is proposed Thealgorithm includes two clustering processes: the first is linear clustering process, based on thematching degree of heuristic clustering process to generate matching degree matrix, throughtransformation of ranks to generate aggregation matrix, and then using roulette strategy todetermine service requirements solely to one vehicle; the second process is called secondaryclustering process, based on the prior clustering algorithm solving SVLRMP problem to realizethe single vehicle carpooling matching process. Considering the probabilistic characteristic oflinear clustering process, in order to improve clustering accuracy and increase coordinationamong vehicles, the active migration (including emigration and immigration process) andperturbation strategy of service demand are put forward. By result analysis of10vehicles and30carpooling service demands in Jinan, China, TSCA algorithm can give good carpooling plans inshort time with high overall carpooling costs and success rate, and present certain vehiclesynergy.3. On the basis of definitive multi-vehicles logistics carpooling matching problem,introduce transferring characteristic and allow one service demand to achieve its goal byriding several vehicles successively, which adds to difficulty of path optimization. This paper tries to simply classify service demands of clients according to relaxation degree of time window,first meeting service demands with high priority and then lowers, trying to find the way toimprove carpooling success rate. Aiming at the service demand transferring phenomenon thatoften appear in the process of logistics scheduling, the concept of transfer is introduced intomulti-vehicle logistics carpooling matching, and the multi-vehicle carpooling matching problemsupporting transfer----MVLRMP-T problem is put forward. The problem discards the constraintthat a service demand can only accept service from one vehicle, which can further improvecarpooling success rate. Definition of the problem is formalized, determination method oftransfer stop is studied and object function and constraint conditions are confirmed. In this paper,the road net structure of MVLRMP-T problem can be formed through optimization conclusion ofMVLRMP and transferring routes can be chosen and confirmed based on improving ant colonyalgorithm. When optimizing carpooling, optimization strategy of first serial optimize, thencombine to fine tune is taken. Experimental results show that improving ant colony algorithm caneffectively support the transfer, presenting strong adaptive ability.4. On the basis of definitive multi-vehicles logistics carpooling matching problem,introduce time-varying characteristic of road net and extend definitive problem to logisticscarpooling matching problem with certain dynamism, to study the influence of temporaland spatial characteristics of road net on vehicle speed and the searching method of pathoptimization under these characteristics. According to the time-varying characteristic of roadnet in actual logistics scheduling process, based on previous study, the multi-vehicle logisticscarpooling matching problem based on temporal and spatial characteristics of road net---MVLRMP-ST problem is proposed. The problem fully considers the temporal and spatialcharacteristics of road net as well as the environmental change situation, being more suitable toactual application. Formalize definition of the MVLRMP-ST problem, slack constraints on timewindow in former MVLRMP problem, and put forward the TSCA-ST algorithm to solve thisproblem. TSCA-ST algorithm is extension of the TSCA algorithm, vehicle’s speed changesaccording to characteristics of road net; determine the virtual travel time, actual travel time andtime frame K of the road segment that vehicle passes, put forward the determination method ofvehicle speed, and consider the influence of environmental characteristics of road net on thequality of the path. To verify validity of the algorithm, this study based on Shandong highwaynetwork and national and provincial network within Shandong province, constructs thetime-varying network for conducting this experiment, and designs relevant experiments. In thisexperiment, both the vehicle type and road characteristics are heterogeneous, and the reliableimpact of environmental changes on vehicle routing is taken into consideration, which makesresults of the experiment have strong use value. We design carpooling experiments of vehicles ondifferent workdays, whose effects are quite good and can provide alternatives for actual logisticscarpooling, having high practical value.