Optimization of regional project portfolio by the cluster analysis method

Abstract

In view of the need to provide specific content to existing regional development strategies, theoretical and practical aspects of the development and implementation of relevant projects are becoming increasingly important. Regional development projects have a number of differences from corporate, which requires the development of specific models for their selection and portfolio optimization. The paper analyzes the main factors influencing the regional project portfolio content. The project initiation procedure is investigated. It is found that the main methods of generating ideas for future projects are brainstorming, expert evaluation and e-mail survey. It is hypothesized that in the case of significant increase in the number of proposed projects, the best solution may be to combine some of them. The project comparison criteria, including goals, project product users, performers, territory, duration, cost and funding sources are identified. Partial similarity criteria for the selected group of projects are designed by the cluster analysis methods. The generalized criterion, which allows concluding about the possibility of project clustering, is singled out among the partial criteria. On the basis of the data obtained, a hierarchical cluster tree, the depth adjustment of which allows the optimum number of projects in the regional portfolio is constructed. Project clustering allows solving several problems: avoiding the dissipation of resources on smaller projects, considering the constructive ideas of the project community representatives; increasing the stakeholders' satisfaction with the results of joint work.