Complex Design Problems Research Articles

Exploring the application of LLM-based AI in UX design: an empirical case study of ChatGPT

ABSTRACT Large language model (LLM) based AI applications are being rapidly adopted by various creativity-related sectors, including UX design. However, as newly emerged applications, how UX designers work with LLMs and how to optimize the benefits of LLMs in UX design remain unclear. Leveraging the widely adopted LLM application ChatGPT as a case study, we recruited 12 experienced frontline UX practitioners from IT companies to use ChatGPT in their everyday design activities for 4 to 6 weeks. Through use logs, questionnaires, and interviews, we collected participants’ usage and reflection data. Our analysis reveals the different functions of ChatGPT in facilitating UX design, such as offering design guidelines, constructing user profiles, and simulating stakeholders. Challenges also emerged, such as limitations in understanding complex design problems and prototyping design ideas. Based on the findings, we reflect on the working dynamics between human workers and AI tools and suggest implications for designing more usable LLM-based AI applications.

Generative design for complex floorplans in high-rise residential buildings: A Monte Carlo tree search-based self-organizing multi-agent system (MCTS-MAS) solution

This research aims to develop an innovative generative design algorithm for complex floorplan layout designs in high-rise residential buildings. Unlike conventional methods, which treat such a task as an optimization problem, we approach it as an adaption problem to be solved by combining a self-organized multi-agent system (MAS) and Monte Carlo tree search (MCTS) algorithm. The approach involves two steps. The first is design constraints pre-processing, which comprises (a) initial user input, and (b) pre-processing for initial setup. The second is building plan generation, involving (a) perception, (b) planning, and (c) action modules. Multiple illustrative examples are explored to verify the proposed generative design algorithm in terms of robustness, efficiency, and scale. We discover that the algorithm can handle complex combinations of input constraints (e.g., number of rooms, room areas, adjacency connections, and building boundaries) and generate plausible floorplans in a robust, efficient way that can be scaled up to bigger, more complex floorplan generative design problems. With proper adaptation, the algorithm can potentially be applied to other building types, such as office or hospital buildings.

Metaheuristic algorithm inspired by enterprise development for global optimization and structural engineering problems with frequency constraints

In this study, a novel metaheuristic optimization algorithm inspired by the enterprise development (ED) process was developed. This process encompasses tasks, structure, technology, and human interactions. A mechanism for switching activities was utilized to determine each step by updating the searched solutions. The ED optimizer underwent evaluation against 50 mathematical functions and 54 IEEE Congress on Evolutionary Computation (CEC) functions, and it was compared with six up-to-date algorithms, three winners of CEC 2020 in real-world single-objective constrained optimization, and ten well-known algorithms. The evaluation results revealed that the ED optimizer outperformed the algorithms in mathematical tests. Using finite element simulation, this novel method was applied to minimize steel structure component weights under various frequency constraints. The engineering designs included a 37-bar planar truss, a 52-bar dome, a 72-bar space truss, a 600-bar dome, and a 1,410-bar dome. Optimization results for these structures showed that the developed ED optimizer yielded the optimal solution and required fewer function evaluations. In conclusion, the ED optimizer is a superior metaheuristic algorithm that can effectively solve complex structural design problems.

CREATION OF A NEW METHODOLOGY FOR THE DEVELOPMENT OF DESIGN COMPETENCE AMONG FUTURE ENGINEERS

This article explores the challenge of developing design competence in engineering students, a skill set increasingly vital in the face of modern technological demands. It introduces a novel methodology aimed at enhancing the training of future engineers, equipping them with the necessary tools to solve complex design and construction problems effectively. Grounded in an in-depth analysis of current engineering education practices, the proposed methodology integrates modern technologies, project-based learning, and an interdisciplinary approach to create a comprehensive and innovative educational framework.

A Cost-Aware Multi-Agent System for Black-Box Design Space Exploration

Abstract Effective coordination of design teams must account for the influence of costs incurred while searching for the best design solutions. This paper introduces a cost-aware multi-agent system (MAS), a theoretical model to 1) explain how individuals in a team should search, assuming that they are all rational utility-maximizing decision-makers, and 2) study the impact of cost on the search performance of both individual agents and the system. First, we develop a new multi-agent Bayesian Optimization framework accounting for information exchange among agents to support their decisions on where to sample in search. Second, we employ a reinforcement learning approach based on the multi-agent deep deterministic policy gradient for training MAS to identify where agents cannot sample due to design constraints. Third, we propose a new cost-aware stopping criterion for each agent to determine when costs outweigh potential gains in search as a criterion to stop. Our results indicate that cost has a more significant impact on MAS communication in complex design problems than in simple ones. For example, when searching in complex design spaces, some agents could initially have low-performance gains, thus stopping prematurely due to negative payoffs, even if those agents could perform better in the later stage of the search. Therefore, global-local communication becomes more critical in such situations for the entire system to converge. The proposed model can serve as a benchmark for empirical studies to quantitatively gauge how humans would rationally make design decisions in a team.

Identifying critical demand scenarios for the robust capacitated network design problem using principal component analysis

AbstractIn this paper, we consider the single‐commodity robust network design problem. Given an undirected graph with capacity installation costs on its edges and a set of scenarios with associated flow balance vectors that represent different scenarios of node supplies and demands, the goal is to find integer edge capacities that minimize the total installation cost and permit a feasible single commodity flow for each scenario. This problem arises, for example, in the design of power networks, which are dimensioned to accommodate many different load scenarios. We propose a new method to identify a small subset of the given scenarios, such that solving the robust network design problem for the smaller scenario set leads to almost the same capacities as solving it for the full scenario set . By considering only the scenarios in , the size of the model that needs to be solved can be reduced substantially, while the error introduced by neglecting the remaining scenarios is kept very small. Our method only employs simple techniques from statistical data analysis, namely principal component analysis (PCA), and convex hull computations in low dimensions. Thus, its computational effort is very small and it is easily applicable to more complex network design problems. We evaluate the effectiveness of the method in computational experiments for instances stemming from offshore power grid planning and telecommunication networks. Our results show that the proposed techniques are indeed well suited to identify small scenario subsets that lead to significantly reduced models with high quality solutions.

Early Risk Quantification Strategy for Design Space Reduction Decisions in Set-Based Design

Perceptions of feasibility in design spaces are susceptible to change if new and conflicting information becomes available later. Design space reduction decisions made in set-based design can amplify vulnerability to new information if remaining design spaces and present perceptions are unable to adapt. This paper considers different ways new information can alter perceptions of feasibility for complex design problems and introduces an early, probabilistic strategy for quantifying the risk of eliminating potential design solutions based on the vulnerability of remaining design spaces to new information. Emergent designs of a set-based design process gauging this risk are evaluated against one neglecting it for an analogous design problem. Early results indicate that the probabilistic model is able to effectively delay design decisions and prevent lock-in while design space understanding is still growing.

The Maximum Attainable Flow and Minimal Cost Problem in a Network

This paper, presents an efficient algorithm that solves such a large class of optimization problems. Ford-Fulkerson determines the maximum flow in a network by iteratively augmenting flow paths until no further improvement is possible. On the other hand, Dijkstra's algorithm excels in finding the shortest path in a weighted graph, making it suitable for minimizing costs in network traversal. However, this paper simultaneously optimizes both objectives (flow and cost) dependently in unique iterations by considering all constraints and objectives holistically. The aim of this work is to develop efficient algorithms that can handle complex optimization problems in transportation, network design, and other fields, ultimately improving resource utilization and minimizing costs as its crucial for enhancing decision-making processes, improving efficiency in resource utilization, and achieving cost savings in diverse applications ranging from transportation networks to production planning. This paper deals about formulating the linear programming for the optimizations problems and finding the maximum amount of flow that can be sent from a source node to a sink node while minimizing the total cost of sending that flow by using simplex method (Two phase method). Through computational experiments and case studies, everybody demonstrate the effectiveness and efficiency of the proposed approach in solving real-world network flow problems. Our method yields efficient algorithms with in smallest numbers of iterations and time that enable the optimal allocation of resources within networks, achieving both maximum flow and minimum cost simultaneously.

Human Designers' Dynamic Confidence and Decision-Making When Working With More Than One Artificial Intelligence

Abstract As artificial intelligence (AI) systems become increasingly capable of performing design tasks, they are expected to be deployed to assist human designers' decision-making in a greater variety of ways. For complex design problems such as those with multiple objectives, one AI may not always perform its expected accuracy due to the complexity of decision-making, and therefore, multiple AIs may be implemented to provide design suggestions. For such assistance to be productive, human designers must develop appropriate confidence in each AI and in themselves and accept or reject AI inputs accordingly. This work conducts a human subjects experiment to examine the development of a human designer's confidence in each AI and self-confidence throughout decision-making assisted by two AIs and how these confidences influence the decision to accept AI inputs. Major findings demonstrate severe decreases in a human designer's confidence especially when working with one or more low-performing AI teammates and/or receiving negative feedback. Additionally, a human designer's decision to accept AI suggestions depends on their self-confidence and confidence in one of the two AIs. Finally, an additional AI does not increase a human designer's likelihood of conforming to AI suggestions. Therefore, in comparison to a scenario with one AI, the results in this work caution against the implementation of an additional AI to AI-assisted decision-making scenarios. The insights also inform the design and management of human–AI teams to improve the outcome of AI-assisted decision-making.

Optimization method of building energy efficiency design based on decomposition multi objective and agent assisted model

In today's global climate crisis, energy-efficient building design is crucial for achieving energy efficiency, environmental sustainability, and resident well-being. However, traditional architectural design is difficult to solve complex multi-objective and multivariate optimization problems. To overcome these challenges, this study proposes a solution: a strategy that combines decomposed multi-objective and agent assisted modeling. The proposed method decomposes complex architectural design problems into multiple manageable sub problems, with each sub problem optimized for a specific design objective. This method effectively simplifies the problem structure, allowing each subproblem to explore its solution space more focused and in-depth. Meanwhile, by combining proxy assisted modeling and utilizing proxy models to approximate actual physical processes or performance evaluations, the computational cost is reduced and the optimization process is accelerated. This study indicates that the improved multi-objective backbone particle swarm optimization algorithm relies on adaptive perturbation factors, with an average measured super volume of 29311 for one bedroom buildings and 49504 for three bedroom buildings. For the same building type, the average volume measurements of the multi-objective particle swarm optimization algorithm assisted by the decomposed surrogate model are 21153 and 40230, respectively. The proposed method effectively addresses complex multi-objective optimization problems in the field of building energy efficiency design, simplifies the problem structure, reduces computational costs through surrogate models, accelerates the optimization process, improves energy efficiency, and can support building construction to better cope with the challenges of climate change.

Research on Space Optimization Design of High-rise Residential Building Based on Genetic Algorithm

Abstract With the rapid development of urbanization and the continuous growth of population, the design and planning of high-rise residential buildings have become increasingly important. The purpose of this study is to explore the space optimization design method of high-rise residential buildings based on genetic algorithm(GA), focusing on the comparative analysis between traditional GA and Adaptive genetic algorithm(AGA). In this paper, AGA is used to establish the spatial optimization model of high-rise residential buildings. By dynamically adjusting the parameters of the algorithm, AGA makes the algorithm better adapt to the characteristics of the problem and improves the search efficiency. The results show that AGA is superior to traditional GA in global convergence probability, especially when the population size is large. AGA improves the adaptability and robustness of the algorithm by dynamically adjusting the crossover and mutation probability. AGA has better flexibility and adaptability in the design of high-rise residential buildings and is expected to provide more optimized solutions for solving complex design problems. The findings of this study provide a useful reference for innovation and sustainable development in the field of high-rise building design and also provide practical methods and tools for the application of GA.

Adaptive surrogate assisted multi-objective optimization approach for highly nonlinear and complex engineering design problems

Despite enormous advances in computer power, computationally costly models impede the use of traditional optimization approaches that must be invoked repeatedly during the optimization process in practical engineering applications. Surrogate models have been found to be a promising endeavor in multi-objective optimization problems involving expensive analysis and simulation processes such as multi-physics modeling and simulation, finite element analysis (FEA), and computational fluid dynamics (CFD. Developing an optimization algorithm that can easily identify the Pareto frontier of highly nonlinear multi-objective optimization problems with less computation cost is the aim of this work. In this paper, an Adaptive Multi-Objective Optimization approach based Surrogate models (AMOS) is developed to reduce computation cost of fitness evaluations and discover the Pareto optima for multi-objective optimization problems with comparable high accuracy. AMOS explores the design space by sampling using LHD to identify promising regions. Then, AMOS exploits the identified promising region by adaptively constructing the most suitable surrogate model, which could be response surface, radial basis, or Kriging surrogates, in the feasible design space based on root mean square error values (RMSE). AMOS stops iterating when a termination criterion is met, and a Pareto frontier is identified based on developed guidance and fitness functions. AMOS has successfully identified the pareto frontier of practical engineering optimization problems with expensive black box functions and significantly reduced the computation cost. The novel method was put to the test utilizing real-world challenges and engineering design examples such vehicle magnetorheological design and wind turbine airfoil geometry.

Requirements, Objectives, Both, or Neither: How to Formulate Complex Design Problems for Innovation Contests

Abstract Technical organizations increasingly rely on innovation contests to find novel ideas for designing complex systems. These activities involve outsiders in the early stages of the design process, leading to ground-breaking designs that often surpass expectations. Here, the contest’s rules document plays a crucial role: this design artifact communicates the organization’s problem and the desired system performance to the participants—significantly impacting the resulting solutions. However, the contest’s nature amplifies the challenges of communicating complex design problems across boundaries. Existing strategies for formulating—i.e., requirement and objective allocation—might not suit this context. We developed an inductive model of their formulation process based on a multiyear field study of five complex innovation contests. We found that a formulation team (or “seeker”) balanced the need to communicate their problem in detail with the risk of excluding valuable participants. Here, they chose among three approaches—incentivize, impose, or subsume—depending on their knowledge of potential solutions and the participants’ capabilities. Notably, the seeker formulated more granularly than the literature describes, employing multiple approaches within each rules document. These findings shed light on a poorly understood aspect of innovation contests, shed new light on a longstanding debate in the engineering design literature, and guide practitioners’ formulation processes.

MeLM, a generative pretrained language modeling framework that solves forward and inverse mechanics problems

We report a flexible multi-modal mechanics language model, MeLM, applied to solve various nonlinear forward and inverse problems, that can deal with a set of instructions, numbers and microstructure data. The framework is applied to various examples including bio-inspired hierarchical honeycomb design, carbon nanotube mechanics, and protein unfolding. In spite of the adaptable nature of the model-which allows us to easily incorporate diverse materials, scales, and mechanical features-it performs well across disparate forward and inverse tasks. Based on an autoregressive attention-model, MeLM effectively represents a large multi-particle system consisting of hundreds of millions of neurons, where the interaction potentials are discovered through graph-forming self-attention mechanisms that are then used to identify relationships from emergent structures, while taking advantage of synergies discovered in the training data. We show that the model can solve complex degenerate mechanics design problems and determine novel material architectures across a range of hierarchical levels, providing an avenue for materials discovery and analysis. To illustrate the use case for broader possibilities, we outline a human-machine interactive MechGPT model, here trained on a set of 1,103 Wikipedia articles related to mechanics, showing how the general framework can be used not only to solve forward and inverse problems but in addition, for complex language tasks like summarization, generation of new research concepts, and knowledge extraction. Looking beyond the demonstrations reported in this paper, we discuss other opportunities in applied mechanics and general considerations about the use of large language models in modeling, design, and analysis that can span a broad spectrum of material properties from mechanical, thermal, optical, to electronic.

Tensor completion algorithm-aided structural color design.

In recent years, structural color has developed rapidly due to its distinct advantages, such as low loss, high spatial resolution and environmental friendliness. Various inverse design methods have been extensively investigated to efficiently design optical structures. However, the optimization method for the inverse design of structural color remains a formidable challenge. Traditional optimization approaches, such as genetic algorithms require time-consuming repetitions of structural simulations. Deep learning-assisted design necessitates prior simulations and large amounts of data, making it less efficient for systems with a small number of features. This study proposes a tensor completion algorithm capable of swiftly and accurately predicting missing datasets based on partially obtained datasets to assist in structural color design. Transforming the complex physical problem of structural color design into a spatial structure relationship problem linking geometric parameters and spectral data. The method utilizes tensor multilinear data analysis to effectively capture the complex relationships associated with geometric parameters and spectral data in higher-order data. Numerical and experimental results demonstrate that the algorithm exhibits high reliability in terms of speed and accuracy for diverse structures, datasets of varying sizes, and different materials, significantly enhancing design efficiency. The proposed algorithm offers a viable solution for inverse design problems involving complex physical systems, thereby introducing a novel approach to the design of photonic devices. Additionally, numerical experiments illustrate that the structural color of cruciform resonators with diamond can overcome the high loss issues observed in traditional dielectric materials within the blue wavelength region and enhance the corrosion resistance of the structure. We achieve a wide color gamut and a high-narrow reflection spectrum nearing 1 by this structure, and the theoretical analysis further verifies that diamond holds great promise in the realm of optics.

A human-machine collaboration frame in daylighting optimization of semi-outdoor space design by using phased synergistic method: A case study

The semi-outdoor space (SOS) of a building is a valuable activity space and a microclimate transition zone for occupants. Current parametric and optimization methods for SOS design limit the flexibility and complexity of the design process and lack human-machine interaction, which can lead to design results that do not meet specific requirements and diverge significantly from designer expectations. Therefore, this paper proposes a new dynamic and in-depth human-machine collaborative design method to solve multi-level, complex semi-outdoor space design problems and provide preferable daylighting performance. A phased synergistic method (PSM) is proposed based on hierarchical iterations of design variables. The new method allows for integrating existing design tools and expands their optimization scope to accommodate early design stages with ambiguous requirements. A case study building in Guangzhou, China, is used to showcase the design process and daylighting performance evaluation results. The case study uses the intuitive design method (IDM) and the non-phased method (NPM) results for performance and usability comparison. The study shows that the new method can find spatial combinations with better daylighting performance. It also illustrates the positive effect of adequate human-machine interaction in enhancing building performance and form optimization.

Educating for complexity in Inclusive Design: From products to service systems

Inclusivity is the explicit and ultimate aim of Inclusive Design. Although the field continues to develop, it remains heavily influenced by product design logic and practices. Drawing on current discussions on how to educate designers for real-world challenges, we argue that Service Design offers a complementary knowledge set that supports novice inclusive designers in engaging with increasingly complex design problems. To this end, we describe how a Service Design workshop was introduced into an ongoing inclusive design project. The participants’ self-reported experiences were captured throughout and results show the workshop’s impact in five key aspects: re-framing the problem-solution space; encouraging a new design logic; challenging a heuristic approach to designing by systematizing the process; shifting views on disability from individual to structural and systemic levels; its effectiveness for student learning. We discuss how these findings have direct implications for building capacity to address increasingly complex design problems and for the future of Inclusive Design education.

Collaborative optimization of composite structure and fiber orientation through real-time correction of deep neural network (DNN) models with elite samples

Gradient-free optimization methods can result in significant computational costs when solving complex structural design problems for composite materials. To this end, this article presents a machine learning-based co-optimization method for composite material structure and fiber orientation. In this approach, DNN are utilized as surrogate models for the optimization problem. Equilibrium optimizer is employed to find real-time optimal solution of the DNN. Subsequently, elite samples are generated based on this optimal solution and used to update the DNN until convergence is achieved. During the post-processing stage, B-spline functions are applied to smooth the density and fiber orientation of the optimized results.

A Novel Two-Stage Methodological Approach for Storage Technology Selection: An Engineering–FAHP–WASPAS Approach

Storage technology selection is a very important design decision that greatly affects the future performance of a warehouse; for example, it greatly affects its costs. In making this decision, the designer is faced with a complex issue. It is necessary to select the appropriate option from a wider set of available technologies, taking into account numerous influencing factors. In design practice, solving this problem is primarily based on the experience of designers and the recommendations of manufacturers of these technologies. In the academic literature, this problem has not been properly posed and solved, so there are no papers that comprehensively address this complex design problem. The main goal of this paper is to fill that gap. The presented approach consists of two basic stages. In the first stage, starting from the definition of the project task, potential technologies are generated and critical factors are considered, in order to arrive at a set of acceptable technologies. In the second stage, these technologies are ranked, and a basis for decision making is created. This stage is based on multi-criteria decision making: the Fuzzy Analytic Hierarchy Process (FAHP) method is used to determine the weights of the criteria, and the Weighted Aggregated Sum Product Assessment (WASPAS) method is used to obtain the rankings. The application of the defined approach is tested on real assignments (distribution warehouse, production warehouse, and holding warehouse) and is proven to be applicable to solving these types of problems. The results obtained for the three tested examples prove the suitability of the application of the proposed approach in terms of the aspects of both the quality of the solution and the speed of obtaining it. Considering the practical application of the suggested and filling the recognized literature gap, evident contributions are achieved.

Inverse design of transmission-type linear-to-circular polarization control metasurface based on deep learning

To solve the time-consuming and complex design problems, the deep learning method is used to realize the inverse predictive design of a transmission-type linear-to-circular polarization control metasurface (TLCPCM). Firstly, the target-generation neural network model (TGNNM) is constructed based on a fully connected neural network. The model selects the critical features of the required electromagnetic performance as design targets, and maps low-dimensional design targets to high-dimensional electromagnetic performance. Secondly, taking the output data of the TGNNM as input data, an inverse-mapping neural network model (IMNNM) is constructed by a convolutional neural network. The prediction performance of the IMNNM is compared with two other inverse-mapping models. The research results show that the IMNNM outperforms the other two networks. Finally, combining TGNNM and IMNNM, four sets of TLCPCM structural parameters are predicted. The research results show that the electromagnetic performances of the metasurface determined by the predicted structural parameters are generally consistent with the given design targets. On this basis, one experimental sample is manufactured. The measurement results are consistent with the simulation results. The research results demonstrate the validity and feasibility of the inverse predictive design method proposed in this paper.