Azahar Tekchand Machwe

Integrating aesthetic criteria with evolutionary processes in complex, free-form design – an initial investigation

This research is a continuation of previous work by the authors relating to the inclusion of aesthetic criteria within an interactive evolutionary design system. The work described extends the design system re the manipulation of a more complex design problem with increased importance placed on aesthetic criterion. The paper initially introduces the previous work before describing more recent research and positioning this in terms of previously published work. Finally, initial results from the further developed design system are presented.

Overcoming representation issues when including aesthetic criteria in evolutionary design

It is well known that for any sort of evolutionary search we must represent the problem solution in a suitable manner since the choice of representation has a large impact on the type and efficiency of the evolutionary search procedure applied. Usually in evolutionary design applications either bit strings or real number parameters are used to encode the problem. However, during the initial design phase a ‘design’ may not be decomposable into real number parameters since the nature of the search space is not well understood and / or the designer wishes to maintain a highly flexible approach whilst establishing an initial configuration. The paper introduces the overall objectives of the project and discusses representation issues before presenting an object-based representation which tries to incorporate the ambiguity present during the initial design phase by working with design elements and objects as members of the chromosome. Ambiguity is particularly acute in this case as the overall project objective is the development of a user-centric evolutionary design system that includes aesthetic criteria evaluation.

Multi-objective analysis of a component-based representation within an interactive evolutionary design system

This article describes research relating to a user-centered evolutionary design system that evaluates both engineering and aesthetic aspects of design solutions during early-stage conceptual design. The experimental system comprises several components relating to user interaction, problem representation, evolutionary search and exploration and online learning. The main focus of the article is the evolutionary aspect of the system when using a single quantitative objective function plus subjective judgment of the user. Additionally, the manner in which the user-interaction aspect affects system output is assessed by comparing Pareto frontiers generated with and without user interaction via a multi-objective evolutionary algorithm (MOEA). A solution clustering component is also introduced and it is shown how this can improve the level of support to the designer when dealing with a complex design problem involving multiple objectives. Supporting results are from the application of the system to the design of urban furniture which, in this case, largely relates to seating design.

Towards an Interactive, Generative Design System: Integrating a `Build and Evolve' Approach with Machine Learning for Complex Freeform Design

The research presented in this paper deals concerns interactive evolutionary design systems and specifically with the Interactive Evolutionary Design Environment (IEDE) developed by the authors. We describe the IEDE concentrating upon the three major components: Component-based Representation; Construction and Repair Agents (providing build and evolve services) and a machine learning sub-system. We also describe the clustering technique utilized within the IEDE to improve the user interactivity of the system.

Reducing user fatigue within an interactive evolutionary design system using clustering and case-based reasoning

User fatigue is a limiting factor in interactive evolutionary design and optimization systems. This work illustrates how user fatigue arising from repetitive evaluations can be reduced by incorporating a case-based machine learning system and by clustering the population. An interactive evolutionary design system for urban furniture design is introduced and used as a test-bed for the implementation. The role of clustering within the system is described and initial results are presented. Results obtained from previous work supporting the choice of a case-based approach to machine learning are then presented and, finally, the results from a multi-user study of the performance of the case-based learning system when applied to the design of urban furniture are included.

User-Centric Evolutionary Computing: Melding Human and Machine Capability to Satisfy Multiple Criteria

This chapter centres around the use of interactive evolutionary computation as a search and exploration tool for open-ended contexts in design. Such contexts are characterized by poor initial definition and uncertainty in terms of objectives, constraints and defining variable parameters. The objective of the research presented is the realization of ‘user-centric’ intelligent systems, i.e., systems which can overcome initial lack of understanding and associated uncertainty, whilst also stimulating innovation and creativity through a high degree of human / machine interaction. Two application areas are used to illustrate how, through the adoption of bespoke visualization techniques, flexible representations, and machine learning agents that ‘observe’ the evolutionary process, this objective can be achieved.

Enabling generative behavior within an interactive evolutionary design system using a component-based representation

The paper describes further research relating to an interactive evolutionary design system (IEDS) with emphasis upon extending it to support free-form design. A generative feedback loop is also introduced which, when using a component based representation, extracts user preferences regarding the various components at run-time. This information is fed back into the system to enhance the quality of the generated solutions. The background work leading to the free- form version of the IEDS for urban furniture design is introduced followed by a description of the component preferences extracted from user evaluated solutions. Finally, results of a multi-user comparative study between randomly generated solutions and solutions generated using preference data is presented.