Work alone then meet: Cognitive attributes and team interaction in engineering design using evolutionary algorithms

Abstract

Engineering products are becoming increasingly more complex and require skills from multiple disciplines to design, something hardly realizable by a single individual. As a result, most organizations form design teams, assembling which, is a challenging task. This work presents an agent-based model along with a framework for investigating team dynamics as the designers work on a representative engineering design problem. The model simulates evolutionary systems wherein, designs are generated, selected, and improved upon, as in biological evolution. A binary-coded genetic algorithm evolves the designs under the influence of the designers’ cognitive attributes, both individually and in a team setting. The results show that product complexity generally decreases the quality of designs, while increasing the number of team members initially increases team performance. However, as the number of team members increases further, their performance levels out and finally declines. An interesting find is that the longer the designers work individually on a design before a team meeting, the better the overall performance. Furthermore, increasing the number of candidate solutions brought to the team meeting initially increases team performance followed by a decrease, indicating that there is an optimal number. The paper concludes with proposed directions that future work can take.