AN ARTIFICIAL NEURAL NETWORK-BASED APPROACH COUPLED WITH TAGUCHI'S METHOD FOR PREDICTING THE TOTAL AVERAGE DURATION OF PROJECTS

Autor: Larbi, Bendada, Mourad, Brioua, Selman, Djeffal, Razi, Morakchi
Rok vydání: 2022
DOI: 10.17605/osf.io/gb53a
Popis: Nowadays, project duration prediction has become of crucial importance for managers since it points out the expectancy-life of project realization. To this end, the Neural Network-based approach coupled with the Taguchi method is used to predict the necessary time, which allows the fulfillment of the targeted project within the prescribed span without delay. Accordingly, the whole process for modeling the targeted problem is described, in which the modeling and simulation of the activities network are introduced for calculating the total average time of project. Then, the neural network approach is adopted to predict the total time for finishing the considered project within the deadlines, where the neural network’s input variables are composed of success probability, improvement and learning factors. While, the output variable is the total average project duration, which is the critical data during design phase. After that, the well-known Taguchi method is purposefully used to optimize the already obtained target by neural network. Finally, Simulation analysis through MATLAB are used to show the efficiency of the proposed approach regarding the workability of the approach when it comes to estimating the deadline of the targeted project.
Databáze: OpenAIRE