New Models for Long Term Multi-Project Scheduling with Resource Levelling

Companies are often faced with the challenge of drawing up a strategic plan for a large number of projects to be carried out over a long-term period. The identification of interdependencies, such as technical conflicts and the need for coordination amongst different parties, imposes a strong interconnectedness amongst thousands of activities across all planned projects. The result is usually an extremely complex scheduling problem, where the goal for the company is to execute its projects with as few conflicts between different activities as possible, while at the same time adhering to budget constraints and achieving an even distribution of their yearly financial expenses.

In this talk I will present and compare mathematical modelling approaches to solve a new multi-objective variant of the resource levelling problem with generalised precedence constraints. I will consider a case study for Banedanmark, the main infrastructure manager for the Danish railway network, who has established a portfolio of a large amount of projects concerning maintenance and upgrade of railway infrastructure across the country to be carried out over the next fifteen years. The current strategic planning tool used by Bandenamark largely requires manual scheduling of individual activities by a human scheduler, and the aim is to instead provide an automated and flexible tool that can produce good schedules in short computation times. Preliminary results based on data provided by Banedanmark will be presented in order to illustrate the effectiveness of the mathematical approaches.

Information:  Dr. M. Schmidt, tel. 82199, e-mail: schmidt2@rsm.nl