scholarly journals Effects of Robust Convex Optimization on Early-Stage Design Space Exploratory Behavior

2020 ◽  
Vol 142 (12) ◽  
Author(s):  
Priya P. Pillai ◽  
Edward Burnell ◽  
Xiqing Wang ◽  
Maria C. Yang
Keyword(s):  
Robust Optimization ◽  
Design Space Exploration ◽  
Design Space ◽  
Early Stage ◽  
Pareto Frontier ◽  
Design Parameters ◽  
Safety Factors ◽  
Worst Case ◽  
Robust Convex Optimization ◽  
Early Stage Design

Abstract Engineers design for an inherently uncertain world. In the early stages of design processes, they commonly account for such uncertainty either by manually choosing a specific worst-case and multiplying uncertain parameters with safety factors or by using Monte Carlo simulations to estimate the probabilistic boundaries in which their design is feasible. The safety factors of this first practice are determined by industry and organizational standards, providing a limited account of uncertainty; the second practice is time intensive, requiring the development of separate testing infrastructure. In theory, robust optimization provides an alternative, allowing set-based conceptualizations of uncertainty to be represented during model development as optimizable design parameters. How these theoretical benefits translate to design practice has not previously been studied. In this work, we analyzed the present use of geometric programs as design models in the aerospace industry to determine the current state-of-the-art, then conducted a human-subjects experiment to investigate how various mathematical representations of uncertainty affect design space exploration. We found that robust optimization led to far more efficient explorations of possible designs with only small differences in an experimental participant’s understanding of their model. Specifically, the Pareto frontier of a typical participant using robust optimization left less performance “on the table” across various levels of risk than the very best frontiers of participants using industry-standard practices.

scholarly journals Early-Stage Uncertainty: Effects of Robust Convex Optimization on Design Exploration

2020 ◽  
Author(s):  
Priya P. Pillai ◽  
Edward Burnell ◽  
Xiqing Wang ◽  
Maria C. Yang
Keyword(s):  
Robust Optimization ◽  
Design Space Exploration ◽  
Early Stage ◽  
Human Subjects ◽  
Model Development ◽  
Pareto Frontier ◽  
Design Parameters ◽  
Safety Factors ◽  
Worst Case ◽  
Robust Convex Optimization

Abstract Engineers design for an inherently uncertain world. In the early stages of design processes, they commonly account for such uncertainty either by manually choosing a specific worst-case and multiplying uncertain parameters with safety factors or by using Monte Carlo simulations to estimate the probabilistic boundaries in which their design is feasible. The safety factors of this first practice are determined by industry and organizational standards, providing a limited account of uncertainty; the second practice is time intensive, requiring the development of separate testing infrastructure. In theory, robust optimization provides an alternative, allowing set based conceptualizations of uncertainty to be represented during model development as optimizable design parameters. How these theoretical benefits translate to design practice has not previously been studied. In this work, we analyzed present use of geometric programs as design models in the aerospace industry to determine the current state-of-the-art, then conducted a human-subjects experiment to investigate how various mathematical representations of uncertainty affect design space exploration. We found that robust optimization led to far more efficient explorations of possible designs with only small differences in an experimental participant’s understanding of their model. Specifically, the Pareto frontier of a typical participant using robust optimization left less performance “on the table” across various levels of risk than the very best frontiers of participants using industry-standard practices.

scholarly journals Evaluation of Static Mapping for Dynamic Space-Shared Multi-task Processing on FPGAs

2021 ◽  
Author(s):  
Umar Ibrahim Minhas ◽  
Roger Woods ◽  
Georgios Karakonstantis
Keyword(s):  
High Performance ◽  
Design Space Exploration ◽  
Design Space ◽  
System Throughput ◽  
Design Parameters ◽  
Temporal Constraints ◽  
Shared Resources ◽  
Task Processing ◽  
High Level ◽  
Performance Computing

AbstractWhilst FPGAs have been used in cloud ecosystems, it is still extremely challenging to achieve high compute density when mapping heterogeneous multi-tasks on shared resources at runtime. This work addresses this by treating the FPGA resource as a service and employing multi-task processing at the high level, design space exploration and static off-line partitioning in order to allow more efficient mapping of heterogeneous tasks onto the FPGA. In addition, a new, comprehensive runtime functional simulator is used to evaluate the effect of various spatial and temporal constraints on both the existing and new approaches when varying system design parameters. A comprehensive suite of real high performance computing tasks was implemented on a Nallatech 385 FPGA card and show that our approach can provide on average 2.9 × and 2.3 × higher system throughput for compute and mixed intensity tasks, while 0.2 × lower for memory intensive tasks due to external memory access latency and bandwidth limitations. The work has been extended by introducing a novel scheduling scheme to enhance temporal utilization of resources when using the proposed approach. Additional results for large queues of mixed intensity tasks (compute and memory) show that the proposed partitioning and scheduling approach can provide higher than 3 × system speedup over previous schemes.

scholarly journals Core-Level Modeling and Frequency Prediction for DSP Applications on FPGAs

2015 ◽  
Vol 2015 ◽  
pp. 1-20
Author(s):  
Gongyu Wang ◽  
Greg Stitt ◽  
Herman Lam ◽  
Alan George
Keyword(s):  
High Performance ◽  
Design Space Exploration ◽  
Design Space ◽  
Space Exploration ◽  
Core Level ◽  
Prediction Methods ◽  
Clock Frequency ◽  
Worst Case ◽  
Model Based ◽  
Dsp Applications

Field-programmable gate arrays (FPGAs) provide a promising technology that can improve performance of many high-performance computing and embedded applications. However, unlike software design tools, the relatively immature state of FPGA tools significantly limits productivity and consequently prevents widespread adoption of the technology. For example, the lengthy design-translate-execute (DTE) process often must be iterated to meet the application requirements. Previous works have enabled model-based, design-space exploration to reduce DTE iterations but are limited by a lack of accurate model-based prediction of key design parameters, the most important of which is clock frequency. In this paper, we present a core-level modeling and design (CMD) methodology that enables modeling of FPGA applications at an abstract level and yet produces accurate predictions of parameters such as clock frequency, resource utilization (i.e., area), and latency. We evaluate CMD’s prediction methods using several high-performance DSP applications on various families of FPGAs and show an average clock-frequency prediction error of 3.6%, with a worst-case error of 20.4%, compared to the best of existing high-level prediction methods, 13.9% average error with 48.2% worst-case error. We also demonstrate how such prediction enables accurate design-space exploration without coding in a hardware-description language (HDL), significantly reducing the total design time.

scholarly journals Biased Information Passing Between Subsystems Over Time in Complex System Design

2015 ◽  
Vol 138 (1) ◽  
Author(s):  
Jesse Austin-Breneman ◽  
Bo Yang Yu ◽  
Maria C. Yang
Keyword(s):  
Complex System ◽  
System Design ◽  
Large Scale ◽  
Early Stage ◽  
Systems Design ◽  
System Level ◽  
Worst Case ◽  
Early Stage Design ◽  
Complex System Design ◽  
Biased Information

During the early stage design of large-scale engineering systems, design teams are challenged to balance a complex set of considerations. The established structured approaches for optimizing complex system designs offer strategies for achieving optimal solutions, but in practice suboptimal system-level results are often reached due to factors such as satisficing, ill-defined problems, or other project constraints. Twelve subsystem and system-level practitioners at a large aerospace organization were interviewed to understand the ways in which they integrate subsystems in their own work. Responses showed subsystem team members often presented conservative, worst-case scenarios to other subsystems when negotiating a tradeoff as a way of hedging against their own future needs. This practice of biased information passing, referred to informally by the practitioners as adding “margins,” is modeled in this paper with a series of optimization simulations. Three “bias” conditions were tested: no bias, a constant bias, and a bias which decreases with time. Results from the simulations show that biased information passing negatively affects both the number of iterations needed and the Pareto optimality of system-level solutions. Results are also compared to the interview responses and highlight several themes with respect to complex system design practice.

scholarly journals Biased Information Passing Between Subsystems Over Time in Complex System Design

2014 ◽  
Author(s):  
Jesse Austin-Breneman ◽  
Bo Yang Yu ◽  
Maria C. Yang
Keyword(s):  
Complex System ◽  
System Design ◽  
Large Scale ◽  
Early Stage ◽  
System Level ◽  
Worst Case ◽  
Team Members ◽  
Early Stage Design ◽  
Complex System Design ◽  
Biased Information

The early stage design of large-scale engineering systems challenges design teams to balance a complex set of considerations. Established structured approaches for optimizing complex system designs offer strategies for achieving optimal solutions, but in practice sub-optimal system-level results are often reached due to factors such as satisficing, ill-defined problems or other project constraints. Twelve sub-system and system-level practitioners at a large aerospace organization were interviewed to understand the ways in which they integrate sub-systems. Responses showed sub-system team members often presented conservative, worst-case scenarios to other sub-systems when negotiating a trade-off as a way of hedging their own future needs. This practice of biased information passing, referred to informally by the practitioners as adding “margins,” is modeled with a series of optimization simulations. Three “bias” conditions were tested: no bias, a constant bias and a bias which decreases with time. Results from the simulations show that biased information passing negatively affects both the number of iterations needed to reach and the Pareto optimality of system-level solutions. Results are also compared to the interview responses and highlight several themes with respect to complex system design practice.

P3GA: An Algorithm for Technology Characterization

2015 ◽  
Vol 137 (1) ◽  
Author(s):  
Edgar Galvan ◽  
Richard J. Malak
Keyword(s):  
Design Automation ◽  
Design Space Exploration ◽  
Design Space ◽  
Space Exploration ◽  
Numerical Test ◽  
Pareto Frontier ◽  
Test Problems ◽  
Genetic Optimization ◽  
Technology Evaluation ◽  
Mathematical Basis

It is important for engineers to understand the capabilities and limitations of the technologies they consider for use in their systems. However, communicating this information can be a challenge. Mathematical characterizations of technical capabilities are of interest as a means to reduce ambiguity in communication and to increase opportunities to utilize design automation methods. The parameterized Pareto frontier (PPF) was introduced in prior work as a mathematical basis for modeling technical capabilities. One advantage of PPFs is that, in many cases, engineers can model a system by composing frontiers of its components. This allows for rapid technology evaluation and design space exploration. However, finding the PPF can be difficult. The contribution of this article is a new algorithm for approximating the PPF, called predictive parameterized Pareto genetic algorithm (P3GA). The proposed algorithm uses concepts and methods from multi-objective genetic optimization and machine learning to generate a discrete approximation of the PPF. If needed, designers can generate a continuous approximation of the frontier by generalizing beyond these data. The algorithm is explained, its performance is analyzed on numerical test problems, and its use is demonstrated on an engineering example. The results of the investigation indicate that P3GA may be effective in practice.

Faster Generation of Feasible Design Points

2005 ◽  
Author(s):  
Bernard Yannou ◽  
Faysal Moreno ◽  
Henri J. Thevenot ◽  
Timothy W. Simpson
Keyword(s):  
Monte Carlo ◽  
Design Space Exploration ◽  
Design Space ◽  
Pareto Frontier ◽  
Solution Space ◽  
Computation Time ◽  
Analytical Form ◽  
Research Field ◽  
Design Under Uncertainty ◽  
Active Research

Design space exploration during conceptual design is an active research field. Most approaches generate a number of feasible design points (complying with the constraints) and apply graphical post-processing to visualize correlations between variables, the Pareto frontier or a preference structure among the design solutions. The generation of feasible design points is often a statistical (Monte Carlo) generation of potential candidates sampled within initial variable domains, followed by a verification of constraint satisfaction, which may become inefficient if the design problem is highly constrained since a majority of candidates that are generated do not belong to the (small) feasible solution space. In this paper, we propose to perform a preliminary analysis with Constraint Programming techniques that are based on interval arithmetic to dramatically prune the solution space before using statistical (Monte Carlo) methods to generate candidates in the design space. This method requires that the constraints are expressed in an analytical form. A case study involving truss design under uncertainty is presented to demonstrate that the computation time for generating a given number of feasible design points is greatly improved using the proposed method. The integration of both techniques provides a flexible mechanism to take successive design refinements into account within a dynamic process of design under uncertainty.

Sign in / Sign up

Export Citation Format

Share Document