Robust multi-criteria project scheduling in plant engineering and construction

Abstract

In this chapter we consider a real-life problem which consists in scheduling the activities of a project subject to precedence and resource constraints so as to optimize several conflicting goals. The durations of the activities cannot be specified precisely in advance. Rather, we assume that based on the experience with previous projects, the means, the standard deviations, and certain percentiles of the respective probability distributions can be reliably estimated. The algorithm applied to solve this problem relies on goal programming techniques in conjunction with Goldratt’s Critical Chain Project Management (CCPM) method. The algorithm was applied to a case study dealing with the construction of an accommodation module for an oil rig. Goal programming is a multi-objective programming technique which attempts to minimize the deviations to a set of target values for the given objectives in such a way that all operational restrictions of the problem are satisfied. Several solutions can be obtained, and the best solution will depend on the priority associated to each goal. In this work we considered the minimization of the project makespan and the levelling of the project resources as the objectives to be pursued. The results obtained using the proposed algorithm have been compared with classical project management techniques (PERT/CPM) that the company involved in the case study used in many projects.