Determining the shortest path between two nodes in a rail transportation network can be a time consuming work, since many trains are moving at the same time and the status of each routes are changing by the movements of the trains. Besides restrictions of the train movements in a railway by using switching point machines requires the concept of super-network.
This thesis states the problem and proposes a best-first search algorithm($A^*$ algorithm) for rail transportation network to minimize the search space and computation time. In this purpose, an evaluation function suitable for $A^*$ algorithm in a super-network is proposed which requires no additional information from the problem domain. Instead of using extra information, a method of assessing the cost of paths is proposed as the evaluation function. Other features required for implementing $A^*$ algorithm in a super-network are also considered. Performance evalution with application examples show that the algorithm is effective in reduction of search space and computation time.