Development of multi-objective optimization model for project portfolio selection using a hybrid method

Abstract

Selecting inappropriate projects and project portfolios can result in irreversible wasted economic opportunities, reduced manpower value, and missed prospects and other resources for the organization. As a result, to achieve the best possible outcome, all criteria to enable the best possible choices to be made should be considered. Choosing projects wisely and managing the project portfolio can assist organizations in gaining a better understanding of their projects and their risks and advantages. When faced with budget and other constraints, the ability to select an optimal mix of projects is a significant advantage in the project selection process. The selection of projects by means of employing an effective method is uncommon because many methods are deemed ineffective due to limitations on the number of projects that can be chosen, along with the failure to select economical projects. Project selection is a complex, multicriteria decision-making procedure involving numerous and frequently competing goals. The complexities of project selection problems stem primarily from the large number of projects that are required to be selected for an appropriate collection of investment projects. The study identified some research gaps such as limited studies on social sustainability benefits, criteria for public project selection not being considered or mentioned, and the decision-making committee or expert generating weight to the deviational variables instead of using weighting techniques. The aim of this study is to employ an integrated approach to establish a multi-objective optimization approach for public project portfolio selection. The specific research objectives are to develop an integrated method of Analytic Hierarchy Process, Goal Programming and Genetic Algorithm (AHP-GP-GA), establish a relationship for the developed models to correct the bias of each model and apply the integrated method in a selected community with a set of projects. Data was collected by compiling a well-structured questionnaire for decision-makers analysed by applying the AHP and GP methods. The composition of the integrated approach includes decision support tool with exact and includes meta-heuristic modelling known as Analytic Hierarchy Process, Goal Programming and Genetic Algorithms (AHP-GP-GA) for solving public project portfolio selection problems. The Analytic Hierarchy Process model was used to develop project selection criteria, assign relative priority weights of decision makers, and determine the overall weight of project alternatives. The GP constructed the mathematical model to handle large numbers of objectives and constraints. The GA is the solution algorithm for the effective and flexible optimization model to produce optimal solutions. The AHP and GA employed Spice Logic and MATLAB software packages to analyse, validate and enhance the research. The AHP model highlighted some sub-criteria and project criteria attributes that are significant to project selection criteria. These criteria are economic development, job creation, community acceptance, structure aligned with company goals, employment record of project manager, locality of the project, finish period of the project selected, project threats and political impact. Meanwhile, empirical research on public agencies was undertaken with the AHP-GP-GA, AHP-GP and GP separately to address the problem. The GP and AHP-GP used the LINGO 18.0 software package, while the developed integrated method AHP-GP-GA was solved using MATLAB software package to exhibit the competence of the model and the research. The high point of the empirical research showed that the AHP-GP-GA model can solve large-scale, or complex problems with a large number of decision variables. It selected more projects compared to the AHP-GP and GP standalone model and provided more optimal solutions, which made the approach robust and flexible for solving decision-making problems. The theoretical and practical contributions of the study are the research, which will improve the knowledge and understanding of researchers or academia in PPSP and add to the literature to enhance the existing methods of integrated approaches. The stakeholders in project management practitioners like organization management, top executives, senior and junior supervisors, and personnel connected to the projects will also benefit from the research in selecting optimal projects from the various solution options, saving costs, and learning how to handle and select more complex projects in large-scale real-life situations. This study recommends further research on the integration of stochastic models, evolutionary algorithms, or computation with AHP and GP for the Public Project Portfolio Selection Problem.