Comparing optimization modeling approaches for the multi-mode resource-constrained multi-project scheduling problem. Issue 4 (11th November 2019)

• Record Type: Journal Article
• Authors: Kannimuthu, Marimuthu; Raphael, Benny; Ekambaram, Palaneeswaran; Kuppuswamy, Ananthanarayanan
• Is Part Of: Engineering, construction and architectural management. Volume 27:Issue 4(2020)
• Page Start: 893
• Page End: 916
• Publication Date: 2019-11-11
• Subjects: Optimization -- Scheduling -- Project management -- Decision support systems -- Construction planning; Construction industry -- Management -- Periodicals; Engineering -- Management -- Periodicals; Engineering -- Periodicals; Building -- Periodicals; 624.068
• DOI: 10.1108/ECAM-03-2019-0156
• Languages: English
• ISSNs: 0969-9988

Abstract

Abstract : Purpose: Construction firms keep minimal resources to maintain productive working capital. Hence, resources are constrained and have to be shared among multiple projects in an organization. Optimal allocation of resources is a key challenge in such situations. Several approaches and heuristics have been proposed for this task. The purpose of this paper is to compare two approaches for multi-mode resource-constrained project scheduling in a multi-project environment. These are the single-project approach (portfolio optimization) and the multi-project approach (each project is optimized individually, and then heuristic rules are used to satisfy the portfolio constraint). Design/methodology/approach: A direct search algorithm called Probabilistic Global Search Lausanne is used for schedule optimization. Multiple solutions are generated that achieve different trade-offs among the three criteria, namely, time, cost and quality. Good compromise solutions among these are identified using a multi-criteria decision making method, Relaxed Restricted Pareto Version 4. The solutions obtained using the single-project and multi-project approaches are compared in order to evaluate their advantages and disadvantages. Data from two sources are used for the evaluation: modified multi-mode resource-constrained project scheduling problem data sets from the project scheduling problem library (PSPLIB) and three real case study projects in India. Findings: Computational results prove the superiority of the single-project approach over heuristic priority rules (multi-project approach). The single-project approach identifies better solutions compared to the multi-project approach. However, the multi-project approach involves fewer optimization variables and is faster in execution. Research limitations/implications: It is feasible to adopt the single-project approach in practice; realistic resource constraints can be incorporated in a multi-objective optimization formulation; and good compromise solutions that achieve acceptable trade-offs among the conflicting objectives can be identified. Originality/value: An integer programming model was developed in this research to optimize the multiple objectives in a multi-project environment considering explicit resource constraints and maximum daily costs constraints. This model was used to compare the performance of the two multi-project environment approaches. Unlike existing work in this area, the model used to predict the quality of activity execution modes is based on data collected from real construction projects.