Use of Artificial Intelligence in a Regulated Design Environment – A Beam Design Example

Abstract

The development of computer design tools in relation to construction management is perceived as a mainstream to automate several construction processes. Automation is not necessarily based on Artificial Intelligence (AI). But if we power it up with data, we can make a machine to automatically mimic and eventually supersede human intelligence. However, until now, not much evident research has been conducted to pursue successful applications of AI in the design. In using AI for design decision making, the intelligent agent should first gather information from the environment and then transform them into internal context. Last, the agent acts on its perceived environment in a way that maximizes its chances of success. To better assess the applicability of AI to the construction management, we classify the design environment into two categories; in a regulated design environment, a collection of written regulations governs the design process. In contrast, the needs and expectations of people are the basis for the input data in a human-controlled design environment. Text analysis is probably the most common way of interpreting design and construction related data. Choosing an appropriate AI technique depends largely on how much we know about the environment and the problem at hand. Search methods and optimization theories seem more suited to solve regulated design problems compared to learning methods. An example of the selection of correct beam size is used to describe the methods for choosing an alternative from a set of acceptable choices (those satisfy the regulations) with the highest possible reward.