Thesis (M.Eng.) - Asian Institute of Technology

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering, School of Advanced Technologies

In this study, the resource constrained project scheduling problem with multiple execution modes is considered, in which an activity can be performed in one out of several modes. Each mode of an activity represents an alternative way of combining different levels of resource requirements with related duration. Resources are grouped into two types: renewable and nonrenewable. A renewable resource is limited in each period of time and becomes available for use again after its current engagement with an activity is over. A non-renewable resource, on the other hand, is also limited in the overall project but its availability gets reduced by the amount of its usage in successive engagements with the activities. The objective of the project scheduling is minimization of the project makespan. Genetic algorithm is applied to solve the problem. The algorithm is based on a precedence based feasible list of activities and mode assignment. The algorithm operator is based on related crossover, mutation and selection. The mathematical models of classical RCPSP and MRCPSP are formulated and programmed in LINGO to get the optimal solution. They are used to compare the performance of the proposed genetic algorithm. Finally, thorough experimentation at various levels of the parameters viz. population size, probability of mutation, and the types of crossovers, the best values are chosen. The results have been tested by using the standard set of instances and are presented as the fraction of average deviation, and maximal deviation from the optimal.