A novel scheduling methodology for resource constrained projects by a new mathematical model and a hybrid metaheuristic: A case study


Basar A.

JOURNAL OF THE FACULTY OF ENGINEERING AND ARCHITECTURE OF GAZI UNIVERSITY, vol.37, no.3, pp.1169-1184, 2022 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 37 Issue: 3
  • Publication Date: 2022
  • Doi Number: 10.17341/gazimmfd.913666
  • Journal Name: JOURNAL OF THE FACULTY OF ENGINEERING AND ARCHITECTURE OF GAZI UNIVERSITY
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, Art Source, Compendex, TR DİZİN (ULAKBİM)
  • Page Numbers: pp.1169-1184
  • Keywords: Project management, Project scheduling, Mathematical modeling, Hybrid metaheuristic, Case study, GENETIC ALGORITHM, ACTIVITY TIMES, OPTIMIZATION
  • Istanbul Technical University Affiliated: Yes

Abstract

Purpose: Although project management techniques have been widely developed in recent years, there is a need for structured methodologies and optimization models for time planning of project activities since traditional methods (i.e., Critical Path Method ??? CPM, Program Evaluation and Review Technique ??? PERT) do not provide optimal results. Moreover, as the parameters and constraints affecting the time planning of projects and the number of activities increase, time management of projects have become more complex. For this reason, this paper presents a new mathematical model to complete activities of resource constrained projects in accordance with their logical relationships. Theory and Methods: Optimal solution of the proposed NP-Hard model cannot be found in case of a high number of activities. Therefore, a hybrid metaheuristic approach based on Tabu Search and Genetic Algorithm is proposed to find efficient solutions for the mathematical model. Moreover, the proposed mathematical model and hybrid metaheuristic are applied to plan the activities of projects in an information technology company in Turkey. In the first step of the application, activities of 154 projects completed in 2018, 2019, and 2020 have been scheduled by the proposed methods. Thus, both the methods have been validated and the best parameters of the proposed techniques (i.e. tabu list size, diversification, stopping criteria, mutation) have been determined. The results found by CPLEX solver and proposed methods have been compared. Since CPLEX cannot provide optimum solutions for the projects with more than 30 activities, the results found in 10 hours have been compared with the proposed methods. In the second step of the application, activities of 64 projects planned to work in 2021 have been scheduled by the proposed methods. Results: It is observed that proposed metaheuristic gives 3.91% better objective function values than CPLEX for 154 projects completed in 2018, 2019, and 2020. Furthermore, proposed methods find these better solutions in a short time (on average 178 seconds) in comparison with CPLEX (on average 34,833 seconds). Moreover, proposed metaheuristic gives 5.01% better results than CPLEX in also very short time (on average 130 seconds) in comparison with CPLEX (on average 35,438 seconds) for 64 projects planned to work in 2021. Conclusion: It is observed that proposed model and solution approach provide efficient solutions with high success ratio and reliable plans. Thus, experts working in the company and the customers have approved the schedules obtained by the proposed methods.