Optimizing resource allocation in a portfolio of projects related to technology infusion using heuristic and meta-heuristic methods

Abstract

In today's competitive environment, manufacturing companies are under constant pressure to improve previous products or release new ones. Nonetheless, most products are not designed and built from scratch, but rather, are based on previous versions of the product with the addition of incremental improvements given by the infusion of new technologies. The objective of this research is to focus on continuous improvements where the level of required change is small to medium, which is the most common manner that companies use to achieve advancements in their products or systems. Most of the available literature related to project scheduling assumes that projects are non-iterative and do not consider rework in the analyzes. On the other hand, studies that analyze cyclical projects focus on product design and development, which usually requires a level of experimentation that makes them inherently different from advancements due to incremental improvements. At the same time, the literature on technology innovation is abundant and there are frameworks to assess the impact of transferring various technologies into existing products. However, there has not been proposed a method that specifically addresses the planning and scheduling process required to infuse technologies. Furthermore, the definitive selection for infusion cannot be applied without taking into account available resources, time required to mature technologies and the interaction among them. Portfolio selection and the scheduling process have usually been treated separately although they are interdependent in this particular case. Different plans can make quite different demands on system resources and its availability will impact the portfolio of selected technologies. This thesis intents to bridge the gap between the portfolio scheduling as well as processes for technology selection and insertion by taking a holistic approach, while the iterative nature of activities, due to rework, is included into the model. Therefore, methods for effectively allocating resources in a portfolio of projects related to technology infusion are recommended. Initially, a heuristic method is proposed based on priority rules. However, as the assumptions of the model are loosened a novel method is suggested that combines Genetic Algorithm (GA) and Artificial Bee Colony (ABC). Numerical results indicate that the hybrid meta-heuristic method based on GA-ABC is effective in finding good resource allocations while considering rework; which is shown, can affect the projects that comprise the portfolio and therefore is worthwhile planning for.