Application of GA optimization for solving precedent relationship problem in project scheduling

Based on Genetic Algorithms (GAs) with the matrix approach, the presented model introduces a pre-decision algorithm combining electronic spreadsheets to visually unfold the precedent relationship of activities in project scheduling problems. Trying to avoid large modification to traditional genetic algorithm operators and chromosome structures, the proposed method is improved by reducing time consumption in searching solutions. To demonstrate the technique, two types of scheduling problems for the construction industry are described and tested: traditional contracts with fixed finishing time constraint and the Build-Operate-Transfer (BOT) contracts which have a flexible finishing time with a reward/penalty arrangement. The results clearly show that the model can generate a promising upgrading when compared with the use of traditional GA method in chromosome structures and searching time. © 2011 Elsevier Ltd.