Cyclic Locomotive Assignment Problem using Ising Mean Field Technique

The locomotive assignment problem is among the most important problems in railway transportation system. It is to assign a set of locomotives of different types to trains in a pre-planned train schedule to provide sufficient power to pull them. A solution for this problem is said to be cyclic if the same set of locomotives arrive to each station which had deported from it during of a specified cycle. There are some works based on operation research algorithms tested on small size of acyclic problem which have disadvantage of large CPU time. Instead we used a feedback neural network based on Ising Mean Field Approach to approximate stochastic simulated annealing with a deterministic process. The planning level of problem is considered and a cyclic solution is presented on the real data extracted from the CN North America railway company. The results were very encouraging for a large problem size; 1622 trains were covered with 866 locomotives in less than 3 minutes.

کلیدواژه ها:

Hopfield Neural Network ، Energy Function ، Penalty Term ، Mean Field Equation ، Train Scheduling ، Locomotive Assignment Problem