Distributed constraint satisfaction problems to model railway scheduling problems

Abstract

Railway Scheduling is considered to be a difficult and time-consuming task. Despite real railway networks being modelled as Constraint Satisfaction Problems (CSPs), they require a huge number of variables and constraints. The general CSP is known to be NP-complete; however, distributed models may reduce the exponential complexity by dividing the problem into a set of subproblems. In this work, we present several proposals to distribute the railway scheduling problem into a set of sub-problems as independently as possible. The first technique carries out a partition over the constraint network, meanwhile the second distributes the problem by trains and the third technique divides the problem by means of contiguous stations.