Multidisciplinary Design Optimization through process integration in the AEC industry: Strategies and challenges

Abstract

Recently Multidisciplinary Design Optimization (MDO) has emerged in the Architecture-Engineering-Construction (AEC) industry to assist designers in making the design process more efficient, by achieving more design alternatives in less time. Currently, MDO is developed with software tools that work together and automatically. However, the technical requisites to develop MDO using Process Integration and Design Optimization (PIDO) platforms are not clearly specified in the design optimization literature. There are many difficulties not covered by the literature: especially the tools’ behavior, and the strategies to deal with PIDO. To determine the technical requirements, the tools’ behavior, the challenges of interoperability and viable strategies, we reviewed the literature and tested five tools. This paper presents the main behavior of the tools we studied, and explains the challenges and strategies to develop MDO through PIDO. We observed three technical tool requisites: component interoperability, tool automation, and model parameterization capabilities. We detected low openness levels of the tool interfaces that did not always enable a full integration with PIDO or permit access to model properties. The scarcity of commands and the presence of pop-up menus impeded performing analyses automatically. Moreover, most of the tools did not allow parametric associations among components, compatibility among themselves or the addition of custom components. The strategies proposed focused on testing the tool interfaces, to validate that each computational process runs automatically, and to confirm that parametric relationships and components are possible. The tools tested were not specifically designed to include full capability to work with PIDO, therefore, enhancements would be needed to meet the three requisites: component interoperability, automation and parameterization. Technological, documentation and programming challenges also emerged when working with tools. We demonstrated that only certain tools can be used with a PIDO platform. However, there may be still other requisites for MDO using different methods that can become the focus of future work.