COMMENTS ON THE USE OF OPTIMISATION TECHNIQUES IN PLANNING FREIGHT TRANSPORT OPERATIONS

Application of computers to the planning of freight operations is a very difficult task because of the necessity to deal simultaneously with a very large number of objects and processes inter-dependent in time and space. Freight transport processes are difficult to define in respect of the great number of descriptive parameters, constraints and non-linear character of cost function. There is a general lack of effective mathematical methods enabling us to define phenomena precisely enough and to solve the optimisation models of medium and short-term planning. There are two general approaches to cope with the problem: 1) to build up purely mathematical models, which can be solved with existing algorithms, disregarding the fact that the results will be far from reality or 2) to build up cybernetic models in a heuristic way, defining reality relatively well by using various optimisation techniques, decision making rules, simulation and common sense. This way supplies us with acceptable solutions, which in general are near to optimum. Over ten years of theoretical and experimental research have brought the author to the conclusion that the second approach is the only one applicable practically to complex problems. The latter part of the paper explains the general concept of a medium and short-term planning system concerning routing, train formation and marshalling yard operations, with related inter-connections between particular subsystems, viz. medium term train routing, train formation, wagon routing, empty wagon distribution, timetabling, daily corrections to the plans, marshalling yard operations ultra short term planning and consequent implications. To illustrate the approach to the problem raised, there follows a description of an experimental system of medium term train routing, train formation and wagon routing plan. This system was tested in 1971-72 on a network of PKP. The system takes into account all the essential factors and constraints concerning trains, lines, stations and yards. A further example is a system of empty wagon distribution dealt with on a weekly and daily basis, with respect to eight hours shifts. Both systems take advantage of the theory of graphs, but to obtain satisfactory results it was necessary to employ additional decision rules and the Monte-Carlo method.

  • Supplemental Notes:
    • Presented at Fourth International Symposium on Railroad Cybernetics, AAR/UIC/IRCA, 21-26 April 1974, Washington, D.C. This paper was also published in the November-December, 1973 issue of Rail International, which is available from E.S.L.
  • Corporate Authors:

    International Union of Railways

    14 rue Jean Rey
    75015 Paris,   France 
  • Authors:
    • Truskolaski, A
  • Publication Date: 1974-4

Media Info

  • Features: Figures;
  • Pagination: 5 p.

Subject/Index Terms

Filing Info

  • Accession Number: 00053778
  • Record Type: Publication
  • Source Agency: International Union of Railways
  • Report/Paper Numbers: Paper
  • Files: TRIS
  • Created Date: Apr 17 1974 12:00AM