Generalised Pattern Search with Restarting Fitness Landscape Analysis

Fitness landscape analysis for optimisation is a technique that involves analysing black-box optimisation problems to extract pieces of information about the problem, which can beneficially inform the design of the optimiser. Thus, the design of the algorithm aims to address the specific features detected during the analysis of the problem. Similarly, the designer aims to understand the behaviour of the algorithm, even though the problem is unknown and the optimisation is performed via a metaheuristic method. Thus, the algorithmic design made using fitness landscape analysis can be seen as an example of explainable AI in the optimisation domain. The present paper proposes a framework that performs fitness landscape analysis and designs a Pattern Search (PS) algorithm on the basis of the results of the analysis. The algorithm is implemented in a restarting fashion: at each restart, the fitness landscape analysis refines the analysis of the problem and updates the pattern matrix used by PS. A computationally efficient implementation is also presented in this study. Numerical results show that the proposed framework clearly outperforms standard PS and another PS implementation based on fitness landscape analysis. Furthermore, the two instances of the proposed framework considered in this study are competitive with popular algorithms present in the literature.

Democracy for Profit?

This chapter evaluates Facebook and Twitter’s algorithms using the framework of democratic space. Prominent critiques highlight their opacity and users’ lack of control; tools like Gobo “fix” these algorithms by increasing their flexibility. But while these solutions might cede more control to individual users, they are insufficient for building democratic communities; the more pressing concern for both Facebook and Twitter is their lack of clear boundaries, which undermines users’ ability to recognize their communities. The chapter concludes by showing how we might “democratize” these algorithms in ways that not only increase user control over their digital environments and the algorithms that structure them, but also help to generate and sustain the communities required to exert that control democratically. Ultimately, the chapter argues that questions of ownership and control must be placed alongside considerations about the communal effects of algorithmic design if we are to build environments supportive of democratic politics.

Algorithmic Design in Hybrid Housing Systems

<p>This thesis utilises digital tools to explore notions of flexibility and resilience in the New Zealand suburban house typology. Through aligning with culturally specific paradigms found in traditional Māori Papakāinga settlements, the research questions current western models of community and connectedness through digital simulations. The methodology brings together social, cultural and climactic forces as key influences to internal domestic programme and overall form. The design process is informed by occupancy requirements associated with family types and projected domestic behaviour. This is mapped to cumulative weather data in relation to location and context. Buildable form is therefore a reflection of site specific conditions and planning in relation to various social configurations influenced by culture and community. A key aspect of this research is the creation of a residential model for multi-generational living. Long term adaptability of this residential model is established through planning for future organic expansion & contraction within the development through the careful consideration of modular building platforms that can deal with varying degrees of social diversity. This design research is largely influenced by pre-Socratic theorists and architects working on translating social, geographical and cultural information into data that can inform computational design and simulations. This form of design interpretation through mathematics has arguably stemmed from the birth of calculus in the 17th century, whereby a formula is used to clarify equations with a multitude of variables often represented by Letters and symbols. Utilizing this knowledge in computer aided design (CAD) allows a designer to produce an equation that represents the process from data to design. Aligning design to the mathematical systems allows the work to represent a quantified, systematic depiction of information as opposed to the romanticized view of the ‘Genius Architect’. The workflow and theory behind this research solidifies the role of algorithmic design in architecture and testing the plausibility of these theories in a housing system. While being largely based on the theories of multi-agent systems and algorithmic design, this system also outlines a modular building technology that embellishes design diversity and flexibility. The architecture proposed utilizes parametric design tools and the concept of housing types in a state of flux, whereby the singular entity of the home is considered as part of a much wider collection of housing situations which is forever changing. By adopting the ecological approach seen in nature we allow the space for intergenerational, bicultural living arrangements that have the flexibility to respond to changes without diminishing the flow of social domains.</p>

Algorithmic Design in Hybrid Housing Systems

<p>This thesis utilises digital tools to explore notions of flexibility and resilience in the New Zealand suburban house typology. Through aligning with culturally specific paradigms found in traditional Māori Papakāinga settlements, the research questions current western models of community and connectedness through digital simulations. The methodology brings together social, cultural and climactic forces as key influences to internal domestic programme and overall form. The design process is informed by occupancy requirements associated with family types and projected domestic behaviour. This is mapped to cumulative weather data in relation to location and context. Buildable form is therefore a reflection of site specific conditions and planning in relation to various social configurations influenced by culture and community. A key aspect of this research is the creation of a residential model for multi-generational living. Long term adaptability of this residential model is established through planning for future organic expansion & contraction within the development through the careful consideration of modular building platforms that can deal with varying degrees of social diversity. This design research is largely influenced by pre-Socratic theorists and architects working on translating social, geographical and cultural information into data that can inform computational design and simulations. This form of design interpretation through mathematics has arguably stemmed from the birth of calculus in the 17th century, whereby a formula is used to clarify equations with a multitude of variables often represented by Letters and symbols. Utilizing this knowledge in computer aided design (CAD) allows a designer to produce an equation that represents the process from data to design. Aligning design to the mathematical systems allows the work to represent a quantified, systematic depiction of information as opposed to the romanticized view of the ‘Genius Architect’. The workflow and theory behind this research solidifies the role of algorithmic design in architecture and testing the plausibility of these theories in a housing system. While being largely based on the theories of multi-agent systems and algorithmic design, this system also outlines a modular building technology that embellishes design diversity and flexibility. The architecture proposed utilizes parametric design tools and the concept of housing types in a state of flux, whereby the singular entity of the home is considered as part of a much wider collection of housing situations which is forever changing. By adopting the ecological approach seen in nature we allow the space for intergenerational, bicultural living arrangements that have the flexibility to respond to changes without diminishing the flow of social domains.</p>

Supporting the use of algorithmic design in architecture: An empirical study of reuse of design knowledge

<p>This thesis tests the reuse of design knowledge as a method to support learning and use of algorithmic design in architecture. The use of algorithmic design systems and programming environments offer architects immense opportunities, providing a powerful means to create geometries and allowing dynamic design exploration, but it can also impose substantial challenges. Architects often struggle with adopting algorithmic design methods (translating a design idea into an algorithm of actions), as well as with the implementation of programming languages, the latter often proving frustrating and creating barriers for both novice and advanced software users. The proposition explored in this thesis is that the reuse of design knowledge can improve architects’ ability to use algorithmic design systems, and reduce the barriers for using programming. This study explores and compares two approaches as a means of accessing and reusing existing design solutions. The first approach is the reuse of abstract algorithmic ‘Design Patterns’. The second is the reuse of algorithmic solutions from specific design cases (Case-Based Design). The research was set up as an experimental comparative study between three test groups: one group using Design Patterns, a second group using Case-Based Design, and the control group. A total of 126 designers participated in the study providing sufficient numbers within each group to permit rigorous studies of the statistical significance of the observed differences. Results of this study illustrate that the systematic inclusion of the Design Patterns approach to the learning strategy of programming in architecture and design, proves to be highly beneficial. The use of abstract solutions improves designers’ ability to overcome programming barriers, and helps architects to adopt algorithmic design methods. The use of Design Patterns also encourages design exploration and experimentation. The use of the Case-Based Design approach seems to be more effective after designers and architects, who are novices in programming, gain more experience with the tool. It encourages more focused reasoning, oriented to the realisation of a particular (originally intended) design outcome. The contribution of this research is to provide empirical evidence that the reuse of abstract and case-based algorithmic solutions can be very beneficial. Results of this study illustrate that both reuse methods can be strategically integrated into design education and architectural practice, supporting learning and use of algorithmic design systems in architecture. The study also identifies potential weaknesses of each approach, proposing areas which could be addressed by future studies.</p>

Supporting the use of algorithmic design in architecture: An empirical study of reuse of design knowledge

<p>This thesis tests the reuse of design knowledge as a method to support learning and use of algorithmic design in architecture. The use of algorithmic design systems and programming environments offer architects immense opportunities, providing a powerful means to create geometries and allowing dynamic design exploration, but it can also impose substantial challenges. Architects often struggle with adopting algorithmic design methods (translating a design idea into an algorithm of actions), as well as with the implementation of programming languages, the latter often proving frustrating and creating barriers for both novice and advanced software users. The proposition explored in this thesis is that the reuse of design knowledge can improve architects’ ability to use algorithmic design systems, and reduce the barriers for using programming. This study explores and compares two approaches as a means of accessing and reusing existing design solutions. The first approach is the reuse of abstract algorithmic ‘Design Patterns’. The second is the reuse of algorithmic solutions from specific design cases (Case-Based Design). The research was set up as an experimental comparative study between three test groups: one group using Design Patterns, a second group using Case-Based Design, and the control group. A total of 126 designers participated in the study providing sufficient numbers within each group to permit rigorous studies of the statistical significance of the observed differences. Results of this study illustrate that the systematic inclusion of the Design Patterns approach to the learning strategy of programming in architecture and design, proves to be highly beneficial. The use of abstract solutions improves designers’ ability to overcome programming barriers, and helps architects to adopt algorithmic design methods. The use of Design Patterns also encourages design exploration and experimentation. The use of the Case-Based Design approach seems to be more effective after designers and architects, who are novices in programming, gain more experience with the tool. It encourages more focused reasoning, oriented to the realisation of a particular (originally intended) design outcome. The contribution of this research is to provide empirical evidence that the reuse of abstract and case-based algorithmic solutions can be very beneficial. Results of this study illustrate that both reuse methods can be strategically integrated into design education and architectural practice, supporting learning and use of algorithmic design systems in architecture. The study also identifies potential weaknesses of each approach, proposing areas which could be addressed by future studies.</p>

Blockchain-based Trusty Buyer Coalition Scheme Using A Group Signature

Without trust, buyers may not join a coalition. Despite the tremendous need for trustworthy relationships in buyer coalitions, no current buyer coalition scheme explicitly tackles confidence issues with blockchain technology. This study proposes an algorithmic design, the blockchain-based trusty buyer coalition scheme, to satisfy the trust requirement among different actors while forming the coalition. All activities forming a coalition through a decentralized public ledger can be explicitly examined. Consequently, the proposed algorithm can ensure anonymity within a community, resulting in trusting relationships. Furthermore, the proposed algorithm can ensure correctness and accountability by recognizing misbehavior and enforcing alternative forms of punishment. Additionally, the discovered algorithm can be applied to mobile commerce applications.