From how much to how many : managing complexity in routine design automation

Abstract

Advances in technology and competitive markets are driving the development of products with shorter times to market. In addition to this, products are becoming more complex, with more functionality, and yet lower prices. This has motivated the development of computer systems that support different phases of the design process, one of which is the synthesis process. This process consists of generating candidate design solutions to design problems. This thesis researches the development of software that supports synthesis processes. This type of software is denoted in this thesis as Computer Aided Synthesis, or CAS. The scope lies within the boundaries of routine design problems of artifacts. In such problems, designers have all knowledge available about the components, parameters, relations and constrains that are used for solving them. Synthesis processes for generating solutions to routine design problems consist of two tasks: (1) generating networks of components and (2) attributing values to unknown parameters. The exact strategy determining how these tasks are performed depends on the distribution of requirements throughout the different levels of detail of the problem, e.g. only on top, only at the bottom or as a mix. However, this relation is not known a priori and is different for different distributions of design requirements. From here that complexity in routine design is attributed to the distribution of design requirements along the different abstractions of the problem. Determining the dependency between design complexity and synthesis strategies is the main challenge of this research. The result is a new complexity management approach: Computational Design Synthesis by Complexity Management. This approach integrates six methods, which have all been developed during this research. The first step of this approach consists in formulating a design problem using the FBS based Formulation method. This method uses FBS representations to aid designers in the process of determining which are the exact components, parameters, relations and constraints playing a role in the design problem. The formal design problem is then decomposed by applying the ADT based VII

Bio-based design methodologies for products, processes, machine tools and production systems

“…Evolutionary approaches in CDS are based on the principles of biological evolution, and encompass a broad range of search techniques, among which the most prevalent are genetic algorithms, evolutionary programming and evolution strategies [25][26][27][28][29][30][31]. The search algorithms of evolutionary approaches are based on genetic adaptive operators, namely, mutation and recombination.…”

“…The biologically inspired design cycle presented in Figure 6 is investigated on a mobile machine tool used as a demonstrator during this investigation. The demonstration cycle is described by various functional, bio-inspired surfaces for different applications such as tool and die making for micro-injection molding, forming processes and others [25,26,27,28]. This may lead to more effective tools for design support.…”

“…Generating numerous candidate design solutions and reducing the population by specifying additional constraints finally leads to a set of design alternatives that are "equally good". When the set becomes small enough, the remaining solutions can be further elaborated in the classical way by human intervention [25][26][27][28][29][30][31][32].…”

Reprint of: Bio-based design methodologies for products, processes, machine tools and production systems

“…Evolutionary approaches in CDS are based on the principles of biological evolution, and encompass a broad range of search techniques, among which the most prevalent are genetic algorithms, evolutionary programming and evolution strategies [25][26][27][28][29][30][31]. The search algorithms of evolutionary approaches are based on genetic adaptive operators, namely, mutation and recombination.…”

“…6 is investigated on a mobile machine tool used as a demonstrator during this investigation. The demonstration cycle is described by various functional, bio-inspired surfaces for different applications such as tool and die making for micro-injection molding, forming processes and others [25,26,27,28]. This may lead to more effective tools for design support.…”