Deep learning has recently been applied to various research areas of design optimization. This study presents the need and effectiveness of adopting deep learning for generative design (or design exploration) research area. This work proposes an artificial intelligent (AI)-based deep generative design framework that is capable of generating numerous design options which are not only aesthetic but also optimized for engineering performance. The proposed framework integrates topology optimization and generative models (e.g., generative adversarial networks (GANs)) in an iterative manner to explore new design options, thus generating a large number of designs starting from limited previous design data. In addition, anomaly detection can evaluate the novelty of generated designs, thus helping designers choose among design options. The 2D wheel design problem is applied as a case study for validation of the proposed framework. The framework manifests better aesthetics, diversity, and robustness of generated designs than previous generative design methods. Nomenclature: expectation : differentiable function of discriminator : differentiable function of generator : noise variable ℒ: loss function of autoencoder c: compliance : density variable ̃: filtered density variable ̃̅ : projected density variable : penalization factor Generative Models for Generative DesignGenerative models, one of the promising deep learning areas, can enhance research on generative design. The generative model is an algorithm for constructing a generator that learns the probability distribution of training data and generates new data based on learned probability distribution. In particular, variational autoencoder (VAE) and generative adversarial network (GAN) are popular generative models used in design optimization, where high-dimensional design variables are encoded in low-dimensional design space [13,14]. In addition, these models are utilized in the design exploration and shape parameterization [8,9].The use of generative model to produce engineering designs directly is limited [23]. However, this study claims that the limitations can be overcome by integrating with topology optimization. First, the generative model requires a number of training data, but accumulated training data for various designs in the industry are confidential and difficult to access. A number of designs obtained from topology optimization are expected to serve as training data. Second, the generative model cannot guarantee feasible engineering. In this case,

Car sharing services promise “green” transportation systems. Two vehicle technologies offer marketable, sustainable sharing: autonomous vehicles (AVs) eliminate customer requirements for car pick-up and return, and battery electric vehicles entail zero emissions. Designing an autonomous electric vehicle (AEV) fleet must account for the relationships among fleet operations, charging station (CS) operations, electric powertrain performance, and consumer demand. This paper presents a system design optimization framework integrating four subsystem problems: fleet size and assignment schedule; number and locations of charging stations; vehicle powertrain requirements; and service fees. We also compare an AEV service and autonomous vehicle (AV) service with gasoline engines. A case study for an autonomous fleet operating in Ann Arbor, MI, is used to examine AEV and AV sharing systems profitability and feasibility for a variety of market scenarios. The results provide practical insights for service system decision makers.

Effective electrification of automotive vehicles requires designing the powertrain's configuration along with sizing its components for a particular vehicle type. Employing planetary gear (PG) systems in hybrid electric vehicle (HEV) powertrain architectures allows various architecture alternatives to be explored, including single-mode architectures that are based on a fixed configuration and multimode architectures that allow switching power flow configuration during vehicle operation. Previous studies have addressed the configuration and sizing problems separately. However, the two problems are coupled and must be optimized together to achieve system optimality. An all-in-one (AIO) system solution approach to the combined problem is not viable due to the high complexity of the resulting optimization problem. This paper presents a partitioning and coordination strategy based on analytical target cascading (ATC) for simultaneous design of powertrain configuration and sizing for given vehicle applications. The capability of the proposed design framework is demonstrated by designing powertrains with one and two PGs for a midsize passenger vehicle.

Governments encourage use of electric vehicles (EV) via regulation and investment to minimize greenhouse gas (GHG) emissions. Manufacturers produce vehicles to maximize profit, given available public infrastructure and government incentives. EV public adoption depends not only on price and vehicle attributes, but also on EV market size and infrastructure available for refueling, such as charging station proximity and recharging length and cost. Earlier studies have shown that government investment can create EV market growth, and that manufacturers and charging station operators must cooperate to achieve overall profitability. This article describes a framework that connects decisions by the three stakeholders (government, EV manufacturer, charging station operator) with preferences of the driving public. The goal is to develop a framework that allows the effect of government investment on the EV market to be quantified. This is illustrated in three scenarios in which we compare optimal public investment for a city in USA (Ann Arbor, Michigan) and one in China (Beijing) to minimize emissions, accounting for customer preferences elicited from surveys conducted in the two countries. Under the modeling assumptions of the framework, we find that high customer sensitivity to prices, combined with manufacturer and charging station operator profit maximization strategies, can render government investment in EV subsidies ineffective, while a collaboration among stakeholders can achieve both emission reduction and profitability. When EV and station designs improve beyond a certain threshold, government investment influence on EV adoption is attenuated apparently due to diminishing customer willingness to buy. Furthermore, our analysis suggests that a diversified government investment portfolio could be especially effective for the Chinese market, with charging costs and price cuts on license plate fees being as important as EV subsidies.

No abstract

No abstract

Recent advances in deep learning enable machines to learn existing designs by themselves and to create new designs. Generative adversarial networks (GANs) are widely used to generate new images and data by unsupervised learning. Certain limitations exist in applying GANs directly to product designs. It requires a large amount of data, produces uneven output quality, and does not guarantee engineering performance. To solve these problems, this paper proposes a design automation process by combining GANs and topology optimization. The suggested process has been applied to the wheel design of automobiles and has shown that an aesthetically superior and technically meaningful design can be automatically generated without human interventions.

Purpose -To solve the trade-offs between marketing and R&D domains and to minimize information loss in new product development (NPD), this study proposes an integrated design process as a new solution to the interface system between the two domains. Design/methodology/approach -House of Quality integrated with multivariate statistical analysis is used for determining important design features. These design features are used as parameters for conjoint analysis and Taguchi method, and then the results of analyses are compared. Sequential application of conjoint analysis and Taguchi method, depending on the differences in utilities and signal to noise ratios, is applied for the integrated design process. An automotive interior design is illustrated for the validation of the integrated design process. Findings -The integrated design process determines a point of compromise between the optimums of conjoint analysis and Taguchi method. Sequential application of two methods ensures full utilization of both methods and no loss of information.Research limitations/implications -More illustrations on NPD are needed to verify the proposed process. Practical implications -The design process suggested in this study can be used for process innovation in six sigma approach and be integrated with value chain intelligently. This study proposes the strategic guideline of the integrated design process for enterprises. Originality/value -The integrated design process suggests the solution for the trade-offs between marketing domain that pursues the utility of product and R&D domain that emphasizes robustness of product quality. This integrated design process will give enterprises competitive advantages in NPD.