The Research Of Some Key Problems On Synthetic Automation Of Marshalling Yard

The work of the marshalling yard consists of many jobs from all the branches. The comperes of all levels not only do the work of planning, forming and harmonizing continually, but also give orders duly and exactly. They do abundant information processing as well as take the affection of dynamic environment and disturb of all stochastic factors into account. So the complexity of the system goes beyond the limit that man can hold and control. So modern information technology, control technology and communication technology must be applied to marshalling yard work, using new transportation organization and operation management model to improve the operating efficiency of stations and comprehensive transportation capacity of marshalling yard.In this doctoral thesis, it analyzes the key problem of the current problems of the synthetic automatization of marshalling yard. It brings forward the overall frame of the synthetic automatization of marshalling yard. It puts emphasis on the study of research methods and ideas, the destination, fundamental, analysis of the information processing, overall structure, function modules of the system of marshalling yard synthetic automatization, and the relation between system of marshalling yard synthetic automatization and intelligentized planning system. It defines that the main problem of the synthetic automatization of marshalling yard is a problem of operation optimization. Then it mainly describes the level relation and affiliation between marshalling yard dispatching system and subsystems such as planning subsystem, phase plan subsystem, dispatching cut plan subsystem.Taking the character of the two-way marshalling yard into account, it brings forward the mathematical model of the shift plan, depicts the destination of shift plan compilation and the constraints of all factors in detail. Then it puts emphasis on the heuristic algorithm which solves the problem of determining the optimal solution of train arriving yard. It effectively resolves the problems of exchanging wagons and algorithm to balance the operation between yards and to decide the optimal train departing yards and wagon flow source. Thus it can guarantee the maximal number of received and sent wagons.On the basis of car flow distributing of shift plan, it depicts the mathematical model of dispatching optimization in phase plan within fine limits and the mathematical model of optimization on leading track using. Also it sets different destination and designs different algorithms to meet the requirements of the real work of marshalling yard, considering the different requirement of real productivity. It designs different algorithms too in accordance with different destination to resolve problems. It accomplished optional algorithm of disconnection order, hump using, organizing order, leading track using, different kinds of dispatching locomotives using as a whole. Then it does some simulating computing, taking the observed data from marshalling yard. It takes full account of the use of locomotives, humps and leading track to make the model answer for reality. At the same time, it uses lots of knowledge and rules when arranging dispatching locomotives. And it can adapt all kinds of different marshalling yards if the rules are changed according with different operation mode of different stations. As a result, it has general commonality.It is for the first time in this thesis that artificial intelligence is used to solve the problem of automatical formation of dispatching cut plan. It mainly resolves the problems of disconnection adapting formation, using rails neatly, segregation restriction of formation, automatical processing of dispatching order and railway telegraph together with empty wagons disposal. It reflects the judging knowledge in the mode of production rules to simulate the process of experts reasoning. Then it implements computer compilation of disconnction and dispatch plan of marshalling yard, especially taking all sorts of constraints about intelligent rail selecting and railway transportation safety into account seriously. It does much work on automatical compilation of disconnection. According with the character of marshalling yard operation, because through trains and sectional trains probably gather in different rails, it changes the optional problem into the problem of forming set off and pick up trains after forming groups by station orders. Thus it completes the compilation of dispatch and formation plan with the least cuts and the best number of dispatch wagons by means of casting down wagon groups, eliminating plans, balancing the number of wagons. At the same time it discusses the problem of automatical compilation of wagons picking-out and placing-in cut plan. It puts forward the algorithm steps of fetching and sending wagons, algorithm flow chart and the methods to implement. Then it makes the compilation of wagons picking-out and placing-in cut plan roboticized.In this thesis it analyzes the essence of the problem of routes selecting and arranging of marshalling yard. Then it brings forward the ideas that the algorithm steps of automatical applying of marshalling yard operations can be divided into two center parts: route distribution and route dispatch. Then it gives the algorithm of access distribution. It builds the base for improving the utilization rate of the stations equipments through constructing equipments network, deciding access between two points, comparing interlock table, finding the points to turn back, distributing the access quickly. It builds the mathematical model of route dispatch, taking the minimal summary of the weighted delayed time of all tasks as the destination, taking the front and back steps of route selecting of the tasks as the dynamic constraint. It gets the solution by means of genetic algorithm. The coding uses the long chromosome whose length is the number of the steps of the operation. Each step is divided into two sections. The orders of the codes decide the beginning time and end time that each step occupies the route and the indicator recurrence. The design is based on the algorithm steps of precedence rule. In order to ensure the feasibility of the solution, it makes the coding rationalized and sorts the operation steps topologically. Then it calculates the beginning time and end time of the operation steps by means of recursion, calculates target values of the codes, which it changes into adaptive values later. It makes choices by combining roulette and optimality. It draws the conclusion that the convergence rate is fast and it can choose the route automatically through the calculations on the actual data.At the end of the thesis it does some in-depth research on an important problem of marshalling yard synthetic automatization, the implementing model of workflow driving. It comes up with the flow of operation of arriving, disconnecting, forming, departing and the operation of through trains. Then it puts forward the driving opportunity of the whole automatization and the logic diagram of realization of workflow with the help of MQ message queue. Thus it makes the synthetic automatization self-driving and improves the level of marshalling yard synthetic automatization of our country.