[Dept. of Computer Science]

Experimentation Project ACS

Title Simulation of job shop scheduling strategies
Student Ruben van der Zwaan
Supervisor Marjan van den Akker, Han Hoogeveen
Related Course(s) Simulation, Scheduling and Timetabling
Description In this project we perform a simulation of the job-shop scheduling problem.
An advanced tabu-search algorithm for this problem (developed by Koen de Bontridder) has shown to be very effective.
In the development of planning and scheduling algorithms, it is common to assume that the data are fixed beforehand, i.e., they are deterministic.
However, in real-world situations, tasks often take longer or shorter than we expected.
To evaluate the impact of this variation (or disturbance), we model the processing times as stochastic variables and perform a discrete-event simulation .
The question is:

Is an advanced algorithm like tabu search still effective when there is variation in the processing times or should we prefer on-line dispatching rules?

The choice of a simulation platform is free. A C++ program with the tabu search algorithm is available to be used in the simulation.

Special Note