Repository of DUTH

Development and evaluation of metaheuristic algorithms for project scheduling (Master thesis)

Μήτσου, Νικόλαος/ Mitsou, Nikolaos

Project scheduling is an important part of project management. In this thesis, we aim to develop a method using artificial intelligence to select the best algorithm for the problem of project scheduling with constrained resources. To achieve this, we developed a neural network that aims to select the most appropriate algorithm for the RCPSP. In order to complete this work and satisfy these objectives, research was carried out that included a search in foreign-language bibliography, scientific articles, as well as scientific works and presentations by third parties. The internet was mainly used to obtain information and textbook notes received in the courses of the postgraduate program. The objectives of the thesis are implemented in four chapters. The first chapter presents the basic concepts of project management and the resource-constrained project scheduling problem (RCPSP). The second chapter presents the literature review conducted to implement the thesis. The third chapter presents the four algorithms that will be used to develop the neural network. Finally, chapter four presents the developed neural network, explains how it works, and describes how the database was developed for its preparation and training.