A Conceptual Framework for Simulating Feedback Loops in Engineering Design

Product development is a business process that organisations use to introduce technological advancements to their products and services. Within this process, engineering design decisions define the architecture of the designed product, which, in turn, governs the structure of the product development process that ensures the quality of the delivered product. Integrating product development and engineering design processes results in a product development system characterised by networks of activities related to each other by feedback loops.
Design iteration and rework are two kinds of feedback loop in product development systems. Design iteration is a form of positive feedback loop that contributes to the quality of the designed product. Rework, on the other hand, is a form of negative feedback loop that increases project duration and cost. Understanding feedback loops in product development is fundamental for effective design management and so the delivery of products on time and within budget.
This research contributes a conceptual framework for (a) understanding and (b) simulating the impact of feedback loops in engineering design. The framework enables the integration of two kinds of simulation model. The first is a discrete event model that reflects the product architecture and its influence on the product development process structure, including potential rework feedback loops. The second is an agent-based model that reflects the social facets of design activity and communication behaviours within design teams including design iteration. In this way, two kinds of feedback loop are captured: rework in the discrete event model and design iteration in the agent-based model. An engineering design case study was used to validate the conceptual framework.
This thesis takes a socio-technical systems perspective on engineering design processes. The conceptual framework includes relevant characteristics for simulating feedback loops in engineering design. The product architecture, which identifies individual parts that need to be designed and infers the product development structure, can be used to derive a design process workflow and so the discrete event model. In parallel, social interactions (actions, states, and behaviours) between designers are used to inform an agent-based simulation model of the design activities for each of the parts in the product architecture. Finally, the framework
defines interaction points between the product development process and design activities which informs interplays between the two kinds of simulation model.