طراحي سيستم هوشمند انتخاب پورتفوليوي پروژه ها با در نظرگيري ارزش در معرض ريسك

DESIGNING AN INTELLIGENT SYSTEM FOR PROJECT PORTFOLIO SELECTION CONSIDERING VAR

در اين نوشتار روشي براي مسيله‌ي انتخاب پروژه‌ها ارايه شده است كه قادر به يك‌پارچه‌سازي ارزيابي پروژه‌هاي منفرد با لحاظ كردن اثرات متقابل آنها روي پورتفوليوي پروژه‌هاي شركت است. همچنين از آن‌جا كه افراد خبره و دانش دروني آنها جزو منابع اصلي و غيرقابل جايگزين يك سازمان بوده و به مرور زمان شكل گرفته‌اند، در اين سيستم از روشي مبتني بر شبكه‌هاي عصبي مصنوعي براي كشف دانش خبرگان استفاده شده است. اين سيستم در يك سازمان به صورت آزمايشي به كار گرفته شده كه خروجي آن نزديك به پورتفوليوي پروژه‌هاي مد نظر مديران سازمان بوده و به متعادل‌سازي ريسك پورتفوليوي پروژه‌ها نيز كمك شاياني كرده است. به منظور اعتبارسنجي روش پيشنهادي، مدل بهينه‌سازي نظير مسيله توسعه داده شده كه علاوه بر كسب نتايج مشابه، زمان حل رويكرد پيشنهادي در مسايل اندازه بزرگ بسيار كم‌تر بوده است.

Project portfolio selection problem has been investigated by many researchers over the decades. This paper presents a method for project selection problem that is able to integrate the evaluation of individual projects by considering their interactions on portfolio of company projects. Moreover, since experts and their knowledge is one of the primary and valuable resources of an organization which evolves over time, the proposed system utilizes an artificial neural network approach to discover the experts' knowledge. This system has been used in a pilot organization in an organization while the output is close to the portfolio of projects considered by the managers of the organization and has also contributed to the project portfolio risk balancing. In order to validate the proposed method, an optimization model similar to the problem has been developed. The proposed approach has obtained similar results to the optimization and also needs much less time to solve large-sized problems using the proposed approach. This study tried to consider the interactions of projects when they are selected simultaneously in the portfolio of projects as well as the use of expert opinions and technical knowledge and experience of the organization in project portfolio selection. The proposed approach has been implemented in a project-oriented organization with varying levels of the number and complexity of the projects that have achieved acceptable results. In order to validate the proposed approach, an optimization model is developed and implemented on three problems (two problems of the case study organization and one instance problem with a large number of projects). The results showed acceptable similarity of results (nearly 90% similarity), but the time taken to solve the proposed approach is far less than the optimization model. The proposed approach is able to solve small, medium, and large problems while it provides timely and reasonable solutions in a reliable manner.