A GENETIC ALGORITHM APPROACH FOR OPTIMIZING CHEMICAL TOWERS CONSTRUCTION PROJECT SCHEDULING WITH DYNAMIC RESOURCES CONSTRAINTS

In this paper we formulate the real-life chemical towers construction project as a dynamic resource constrained project scheduling problem (DRCPSP) with minimum makespan. Unlike the traditional resource constrained project scheduling problem, the DRCPSP model is able to consider both independent and dynamic resources which depend on the welding jobs processing sequence (JPS) of the major cylindrical columns components. A modified genetic algorithm with auto-shift mechanism (GAASM) is proposed to search for the optimal solution. A real-life example is presented to demonstrate the applicability of GAASM as well. Simulation experiments of eight illustrative problems with 30 runs show that GAASM outperforms the conventional GA based method (GABM). Furthermore, from eight groups with a total of 240 problems comparing two common rules, namely Top-down and Bottom-up rules, adopted in the current company plans (CP), GAASM demonstrates 5.09% similar to 14.60% average reduction in makespan successfully. Significance: In the DRCPSP model, the resources characteristics are not only independent but also dynamic, which depend on the processing sequence of jobs. The proposed GAASM solution method of DRCPSP can be applied in practice for optimizing chemical towers construction project scheduling.