Crew scheduling for an Urban transit company

Abstract

After a presentation of the problem: crew scheduling for drivers in an urban transit company, we exhibit the difficulties of this kind of optimization problem.

Afterward, an algorithm which is a coupling between a stochastic (annealing-like) algorithm and linear programming is proposed to solve it.

This algorithm is used to constitute the timetable for bus drivers of the T2C company (buses of the city of Clermont-Ferrand) and some of the obtained results are given.

To conclude, the efficiency of this approach is demonstrated by comparisons with various softwares proposed to the T2C by specialized companies. Moreover the structure of our algorithms allows to solve many other kinds of timetable problems. In any urban transit company, to answer the demand, and to improve the working conditions, and to maximize its profit (minimizing its manpower), appears highly combinatoric optimization problems. Many constraints need to be satisfied, a lot of them being “soft” constraints. The combinatoric of the problem is not easily reduced by these constraints. The cost function becomes usually vague and it is no more necessary to find the best solution, but only a good one.