Multi-Objective Optimisation of Urban Form: A Framework for Selecting the Optimal Solution

Abstract

The complexity associated with the design of urban tissues is driven by the multitude of design goals that influence urban development and growth. This complexity is amplified by the design goals being inherently conflicting, necessitating preference-based decisions within the design process—an approach that results in predetermined design solutions driven by personal biases. The utility of population-based optimisation algorithms addresses this by allowing for the examination of multiple conflicting objectives within the same design problem, negating the need for trade-off decisions between the design goals. The application of these algorithms is associated with three primary steps. The first is the formulation of the design problem, the second is the application of the algorithm, and the third is selecting the most optimal solution from the algorithm’s output. This paper examines the third step in this process, in which various methods are employed to facilitate data-driven selection mechanisms that are both objective as well as subjective in their formulation. The selection mechanisms are demonstrated on a speculative urban tissue that examines the potential of inhabiting interstitial spaces, through various morphological interventions, within the urban fabric. The results present a scalable and adaptable framework that assists designers employing multi-objective evolutionary algorithms (MOEAs) to select the optimal solution from their generated populations, a challenge commonly associated with the application of MOEAs in design.

Urban flood resilience in Kampung Melayu: A multi-objective evolutionary approach

“…Unlike a conventional design approach that employs preference based decision making in the early stages of the design process, utilising a MOEA allows the designer to objectively design for the design goals with minimal preference, thus allowing for greater opportunity to explore design options otherwise unattainable through a preference based approach. Although highly beneficial in optimising for multiple conflicting objectives, one of the challenges associated with the application of a MOEA is the filtration of a population of solutions to a single selected solution, this challenge will be addressed in depth in the presented experiment, building on recent selection methods developed by Showkatbakhsh and Makki 20 which highlight the importance of quantitative urban analysis to differentiate between the algorithm’s results.…”

Contextualized computations: a multi-objective optimization approach for designing contextually responsive building envelopes

“…The process of identifying the optimal solution in multi-objective optimization is a complex task. It requires a clear understanding of the designer's requirements and preferences to effectively identify an optimized solution (Showkatbakhsh and Makki, 2022). The results were first examined in terms of the solutions that provided the best performance in a single fitness function.…”

Multi-performative façade systems: The case of real-time adaptive BIPV shading systems to enhance energy generation potential and visual comfort

“…One of the main advantages of an MOEA is that the designer can integrate multiple conflicting fitness functions to evaluate each solution simultaneously, thus allowing the algorithm to evolve a population of solutions that have been independently optimized to the different fitness functions (through the continuous minor improvement of solutions through incremental mutations), consequently generating a varied population of optimized phenotypes. MOEAs have been used extensively across multiple disciplines since the mid-20th century; with a sharp increase in their use within Architecture and Design in the past decade due to the proliferation of various MOEA tools within mainstream 3d modeling software (Showkatbakhsh and Mohammed, 2022).…”