Constraint Reordering for Multi-Objective Configuration Design

1998 ◽  
Author(s):  
Nathan J. Adams ◽  
Georges M. Fadel
Keyword(s):  
Genetic Algorithms ◽  
Design Problem ◽  
Iterative Procedure ◽  
Complex Configuration ◽  
Configuration Design ◽  
Spatial Constraints ◽  
Design Problems ◽  
Acceptable Solution ◽  
The Many ◽  
Combinatorial Methods

Abstract Configuration design is the process of placing components, without altering their shape or connectivity, into an available space, while satisfying various spatial constraints, such as no component overlap. Minimizing the volume occupied by the components and or maximizing the accessibility of the components are just two examples of the many objectives that can drive a configuration design problem. For complex configuration designs, there can be many objectives, which can impose spatial constraints among the components and increase the design complexity, cycle cost, and time. An iterative procedure becomes necessary to reconcile these spatial constraints. To reach solutions that are optimal, these constraints must be reordered else combinatorial methods such as Genetic Algorithms that are used for such problems do not converge. Successful reordering can make complex configuration design problems easier to solve by minimizing the iterations necessary to reach an acceptable solution. Minimizing iterations translates into faster convergence and thus savings on time and money. This paper presents a methodology that can manage the propagation of spatial constraints in complex configuration design problems. Representative examples are shown and results and conclusions are drawn.

Using Q-Learning and Genetic Algorithms to Improve the Efficiency of Weight Adjustments for Optimal Control and Design Problems

2005 ◽  
Author(s):  
Kaivan Kamali ◽  
Lijun Jiang ◽  
John Yen ◽  
K. W. Wang
Keyword(s):  
Optimal Control ◽  
Genetic Algorithms ◽  
Convergence Rate ◽  
Cost Function ◽  
Design Problem ◽  
Numerical Experiments ◽  
Design Parameters ◽  
Design Problems ◽  
Q Learning ◽  
Layered Approach

In traditional optimal control and design problems, the control gains and design parameters are usually derived to minimize a cost function reflecting the system performance and control effort. One major challenge of such approaches is the selection of weighting matrices in the cost function, which are usually determined via trial and error and human intuition. While various techniques have been proposed to automate the weight selection process, they either can not address complex design problems or suffer from slow convergence rate and high computational costs. We propose a layered approach based on Q-learning, a reinforcement learning technique, on top of genetic algorithms (GA) to determine the best weightings for optimal control and design problems. The layered approach allows for reuse of knowledge. Knowledge obtained via Q-learning in a design problem can be used to speed up the convergence rate of a similar design problem. Moreover, the layered approach allows for solving optimizations that cannot be solved by GA alone. To test the proposed method, we perform numerical experiments on a sample active-passive hybrid vibration control problem, namely adaptive structures with active-passive hybrid piezoelectric networks (APPN). These numerical experiments show that the proposed Q-learning scheme is a promising approach for.

Using Q-Learning and Genetic Algorithms to Improve the Efficiency of Weight Adjustments for Optimal Control and Design Problems

2007 ◽  
Vol 7 (4) ◽  
pp. 302-308 ◽  
Author(s):  
Kaivan Kamali ◽  
L. J. Jiang ◽  
John Yen ◽  
K. W. Wang
Keyword(s):  
Optimal Control ◽  
Genetic Algorithms ◽  
Convergence Rate ◽  
Cost Function ◽  
Design Problem ◽  
Numerical Experiments ◽  
Design Problems ◽  
Q Learning ◽  
Complex Design ◽  
Layered Approach

In traditional optimal control and design problems, the control gains and design parameters are usually derived to minimize a cost function reflecting the system performance and control effort. One major challenge of such approaches is the selection of weighting matrices in the cost function, which are usually determined via trial-and-error and human intuition. While various techniques have been proposed to automate the weight selection process, they either can not address complex design problems or suffer from slow convergence rate and high computational costs. We propose a layered approach based on Q-learning, a reinforcement learning technique, on top of genetic algorithms (GA) to determine the best weightings for optimal control and design problems. The layered approach allows for reuse of knowledge. Knowledge obtained via Q-learning in a design problem can be used to speed up the convergence rate of a similar design problem. Moreover, the layered approach allows for solving optimizations that cannot be solved by GA alone. To test the proposed method, we perform numerical experiments on a sample active-passive hybrid vibration control problem, namely adaptive structures with active-passive hybrid piezoelectric networks. These numerical experiments show that the proposed Q-learning scheme is a promising approach for automation of weight selection for complex design problems.

Optimal Designs by Means of Genetic Algorithms

2021 ◽  
pp. 344-354
Author(s):  
Lata Nautiyal ◽  
Preeti Shivach ◽  
Mangey Ram
Keyword(s):  
Genetic Algorithms ◽  
Modeling And Simulation ◽  
Engineering Design ◽  
Design Problem ◽  
High Performance ◽  
Mechanical Engineering ◽  
Design Parameters ◽  
Design Problems ◽  
Simulation Tools ◽  
Engineering Design Problems

With the advancement in contemporary computational and modeling skills, engineering design completely depends upon on variety of computer modeling and simulation tools to hasten the design cycles and decrease the overall budget. The most difficult design problem will include various design parameters along with the tables. Finding out the design space and ultimate solutions to those problems are still biggest challenges for the area of complex systems. This chapter is all about suggesting the use of Genetic Algorithms to enhance maximum engineering design problems. The chapter recommended that Genetic Algorithms are highly useful to increase the High-Performance Areas for Engineering Design. This chapter is established to use Genetic Algorithms to large number of design areas and delivered a comprehensive conversation on the use, scope and its applications in mechanical engineering.

Optimal Designs by Means of Genetic Algorithms

2018 ◽  
pp. 151-161
Author(s):  
Lata Nautiyal ◽  
Preeti Shivach ◽  
Mangey Ram
Keyword(s):  
Genetic Algorithms ◽  
Modeling And Simulation ◽  
Engineering Design ◽  
Design Problem ◽  
High Performance ◽  
Mechanical Engineering ◽  
Design Parameters ◽  
Design Problems ◽  
Simulation Tools ◽  
Engineering Design Problems

With the advancement in contemporary computational and modeling skills, engineering design completely depends upon on variety of computer modeling and simulation tools to hasten the design cycles and decrease the overall budget. The most difficult design problem will include various design parameters along with the tables. Finding out the design space and ultimate solutions to those problems are still biggest challenges for the area of complex systems. This chapter is all about suggesting the use of Genetic Algorithms to enhance maximum engineering design problems. The chapter recommended that Genetic Algorithms are highly useful to increase the High-Performance Areas for Engineering Design. This chapter is established to use Genetic Algorithms to large number of design areas and delivered a comprehensive conversation on the use, scope and its applications in mechanical engineering.

A computational market model for distributed configuration design

1995 ◽  
Vol 9 (2) ◽  
pp. 125-133 ◽  
Author(s):  
Michael P. Wellman
Keyword(s):  
Design Problem ◽  
Original Problem ◽  
Market Model ◽  
Optimal Designs ◽  
Configuration Design ◽  
Distributed Design ◽  
Design Framework ◽  
Design Problems

AbstractA precise market model for a well-defined class of distributed configuration design problems is presented. Given a design problem, the model defines a computational economy to allocate basic resources to agents participating in the design. The result of running these “design economies” constitutes the market solution to the original problem. After defining the configuration design framework, the mapping to computational economies and the results to date are described. For some simple examples, the system can produce good designs relatively quickly. However, analysis shows that the design economies are not guaranteed to find optimal designs, and some of the major pitfalls are identified and discussed. Despite known shortcomings and limited explorations thus far, the market model offers a useful conceptual viewpoint for analyzing distributed design problems.

The Nesting of Two-Dimensional Shapes Using Genetic Algorithms

1995 ◽  
Vol 209 (2) ◽  
pp. 115-124 ◽  
Author(s):  
H S Ismail ◽  
K K B Hon
Keyword(s):  
Genetic Algorithms ◽  
Evolutionary Process ◽  
Search Space ◽  
Cutting Stock ◽  
Two Dimensional ◽  
Optimal Solutions ◽  
Spatial Constraints ◽  
Global Optimal ◽  
Optimum Solution ◽  
Rapid Conversion

The general two-dimensional cutting stock problem is concerned with the optimum layout and arrangement of two-dimensional shapes within the spatial constraints imposed by the cutting stock. The main objective is to maximize the utilization of the cutting stock material. This paper presents some of the results obtained from applying a combination of genetic algorithms and heuristic approaches to the nesting of dissimilar shapes. Genetic algorithms are stochastically based optimization approaches which mimic nature's evolutionary process in finding global optimal solutions in a large search space. The paper discusses the method by which the problem is defined and represented for analysis and introduces a number of new problem-specific genetic algorithm operators that aid in the rapid conversion to an optimum solution.

Design Option Decision Tree: A Method for Schematic Analysis of a Design Problem and Integration of Human Factors Data

1974 ◽  
Vol 18 (3) ◽  
pp. 368-375
Author(s):  
William B. Askren ◽  
Kenneth D. Korkan
Keyword(s):  
Decision Tree ◽  
Human Factors ◽  
Design Problem ◽  
Aircraft Design ◽  
Decision Point ◽  
Specific Design ◽  
Design Problems ◽  
Design Option ◽  
Design Options ◽  
Evaluation Parameters

A Design Option Decision Tree (DODT) is a graphic means of showing the design options available at each decision point in the design process. Several examples of DODTs for aircraft design problems are shown. The procedures for developing a DODT are described. A proposed method for use of the DODT to resolve a design problem is presented. This method includes evaluating the design options in the Tree for impact on the system, and tracing paths through the Tree as dictated by specific design goals. The use of human factors data as one of the evaluation parameters is illustrated. The paper concludes with a discussion of other uses of a DODT.

scholarly journals Students as Researchers, Collaborators, & End-users: Informing Campus Design

2017 ◽  
Vol 8 (2) ◽  
Author(s):  
Amy M. Huber ◽  
Lisa K. Waxman ◽  
Stephanie Clemons
Keyword(s):  
Research Methods ◽  
Interior Design ◽  
Design Problem ◽  
Original Research ◽  
Design Problems ◽  
Campus Design ◽  
Job Opportunities ◽  
Apply Research ◽  
Research Findings ◽  
Integration Strategies

Students in undergraduate design programs often lack opportunity to conduct original research and apply their findings to project solutions. Consequently, they struggle with identifying and framing a design problem, understanding the importance of research-based design, and how to appropriately apply research findings to the needs and desires of project stakeholders. In interior design, this unawareness can lead to design solutions that appeal to the eye, but lack defensible rationale and often do not solve the design problem, or meet user needs. Exposure to research methods and collaborations with practitioners may change how students approach design problems by fostering an empathetic understanding of the human experience.This design case describes a project design at two universities where 72 sophomore and junior students collaborated with furniture manufacturer Herman Miller, Inc. to generate original research before applying their findings to the redesign of informal learning spaces in their campus libraries. Constructivist Learning and Backward Instructional Design, guided the design of the project. The result of this engagement, exposed students to research methods and research integration strategies, who outwardly demonstrated more confidence in making decisions during the design process. While the long-term implications from this type of engagement are not yet evident, encouraging students to ground their design ideas on evidence they have gathered, and their analysis of it, may not only shape their future decision making, but potentially lead to more appropriate client solutions and provide students with coveted job opportunities in positions where evidence-based design is highly valued.

scholarly journals Genetic algorithms in computer aided inductor design

2004 ◽  
Vol 23 (3) ◽  
pp. 68-78
Author(s):  
Jean Fivaz ◽  
Willem A. Cronjé
Keyword(s):  
Genetic Algorithms ◽  
Power Electronics ◽  
Computer Aided Design ◽  
Design Tool ◽  
Design Problems ◽  
Characteristic Data ◽  
Computer Aided ◽  
Aided Design

The goal of this investigation is to determine the advantages of using genetic algorithms in computer-aided design as applied to inductors.  These advantages are exploited in design problems with a number of specifications and constraints, as encountered in power electronics during practical inductor design. The design tool should be able to select components, such as cores and wires, from databases of available components, and evaluate these choices based on the components’ characteristic data read from a database of manufacturers’ data-sheets.  The proposed design must always be practically realizable, as close to the desired specifications as possible and within any specified constraints.

Sign in / Sign up

Export Citation Format

Share Document