skip to main content

This content will become publicly available on January 1, 2023

Title: Plausible Screening Using Functional Properties for Simulations with Large Solution Spaces
When working with models that allow for many candidate solutions, simulation practitioners can benefit from screening out unacceptable solutions in a statistically controlled way. However, for large solution spaces, estimating the performance of all solutions through simulation can prove impractical. We propose a statistical framework for screening solutions even when only a relatively small subset of them is simulated. Our framework derives its superiority over exhaustive screening approaches by leveraging available properties of the function that describes the performance of solutions. The framework is designed to work with a wide variety of available functional information and provides guarantees on both the confidence and consistency of the resulting screening inference. We provide explicit formulations for the properties of convexity and Lipschitz continuity and show through numerical examples that our procedures can efficiently screen out many unacceptable solutions.
Authors:
; ;
Award ID(s):
1854562 1953111
Publication Date:
NSF-PAR ID:
10335111
Journal Name:
Operations Research
ISSN:
0030-364X
Sponsoring Org:
National Science Foundation
More Like this
  1. The world faces an increasing need to phase out harmful chemicals and design sustainable alternatives across various consumer products and industrial applications. Alternatives assessment is an emerging field with focus on identifying viable solutions to substitute harmful chemicals. However, current methods fail to consider trade-offs from human and ecosystem exposures, and from impacts associated with chemical supply chains and product life cycles. To close this gap, we propose a life cycle based alternatives assessment (LCAA) framework for consistently integrating quantitative exposure and life cycle impact performance in the substitution process. We start with a pre-screening based on function-related decision rules,more »followed by three progressive tiers from (1) rapid risk screening of various alternatives for the consumer use stage, to (2) an assessment of chemical supply chain impacts for selected alternatives with substantially different synthesis routes, and (3) an assessment of product life cycle impacts for alternatives with substantially different product life cycles. Each tier focuses on relevant impacts and uses streamlined assessment methods. While the initial risk screening will be sufficient for evaluating chemicals with similar supply chains, each additional tier helps further restricting the number of viable solutions, while avoiding unacceptable trade-offs. We test our LCAA framework in a proof-of-concept case study for identifying suitable alternatives to a harmful plasticizer in household flooring. Results show that the use stage dominates human health impacts across alternatives, supporting that a rapid risk screening is sufficient unless very different supply chains or a broader set of alternative materials or technologies are considered. Combined with currently used indicators for technical and economic performance, our LCAA framework is suitable for informing function-based substitution at the level of chemicals, materials and product applications to foster green and sustainable chemistry solutions.« less
  2. We propose a framework and specific algorithms for screening a large (perhaps countably infinite) spaceof feasible solutions to generate a subset containing the optimal solution with high confidence. We attainthis goal even when only a small fraction of the feasible solutions are simulated. To accomplish it weexploit structural information about the space of functions within which the true objective function lies, andthen assess how compatible optimality is for each feasible solution with respect to the observed simulation outputs and the assumed function space. The result is a set of plausible optima. This approach can be viewed as a way tomore »avoid slow simulation by leveraging fast optimization. Explicit formulations of the general approach are provided when the space of functions is either Lipschitz or convex. We establish both small- and large-sample properties of the approach, and provide two numerical examples.« less
  3. Successful HPC software applications are long-lived. When ported across machines and their compilers, these applications often produce different numerical results, many of which are unacceptable. Such variability is also a concern while optimizing the code more aggressively to gain performance. Efficient tools that help locate the program units (files and functions) within which most of the variability occurs are badly needed, both to plan for code ports and to root-cause errors due to variability when they happen in the field. In this work, we offer an enhanced version of the open-source testing framework FLiT to serve these roles. Key newmore »features of FLiT include a suite of bisection algorithms that help locate the root causes of variability. Another added feature allows an analysis of the tradeoffs between performance and the degree of variability. Our new contributions also include a collection of case studies. Results on the MFEM finite-element library include variability/performance tradeoffs, and the identification of a (hitherto unknown) abnormal level of result-variability even under mild compiler optimizations. Results from studying the Laghos proxy application include identifying a significantly divergent floating-point result-variability and successful root-causing down to the problematic function over as little as 14 program executions. Finally, in an evaluation of 4,376 controlled injections of floating-point perturbations on the LULESH proxy application, we showed that the FLiT framework has 100% precision and recall in discovering the file and function locations of the injections all within an average of only 15 program executions.« less
  4. Abstract

    Accurate and efficient predictions of the quasiparticle properties of complex materials remain a major challenge due to the convergence issue and the unfavorable scaling of the computational cost with respect to the system size. QuasiparticleGWcalculations for two-dimensional (2D) materials are especially difficult. The unusual analytical behaviors of the dielectric screening and the electron self-energy of 2D materials make the conventional Brillouin zone (BZ) integration approach rather inefficient and require an extremely densek-grid to properly converge the calculated quasiparticle energies. In this work, we present a combined nonuniform subsampling and analytical integration method that can drastically improve the efficiency ofmore »the BZ integration in 2DGWcalculations. Our work is distinguished from previous work in that, instead of focusing on the intricate dielectric matrix or the screened Coulomb interaction matrix, we exploit the analytical behavior of various terms of the convolved self-energy Σ(q) in the smallqlimit. This method, when combined with another acceleratedGWmethod that we developed recently, can drastically speed up (by over three orders of magnitude)GWcalculations for 2D materials. Our method allows fully convergedGWcalculations for complex 2D systems at a fraction of computational cost, facilitating future high throughput screening of the quasiparticle properties of 2D semiconductors for various applications. To demonstrate the capability and performance of our new method, we have carried out fully convergedGWcalculations for monolayer C2N, a recently discovered 2D material with a large unit cell, and investigate its quasiparticle band structure in detail.

    « less
  5. ABSTRACT Transcriptional reporters are common tools for analyzing either the transcription of a gene of interest or the activity of a specific transcriptional regulator. Unfortunately, the latter application has the shortcoming that native promoters did not evolve as optimal readouts for the activity of a particular regulator. We sought to synthesize an optimized transcriptional reporter for assessing PhoB activity, aiming for maximal “on” expression when PhoB is active, minimal background in the “off” state, and no control elements for other regulators. We designed specific sequences for promoter elements with appropriately spaced PhoB-binding sites, and at 19 additional intervening nucleotide positionsmore »for which we did not predict sequence-specific effects, the bases were randomized. Eighty-three such constructs were screened in Vibrio fischeri , enabling us to identify bases at particular randomized positions that significantly correlated with high-level “on” or low-level “off” expression. A second round of promoter design rationally constrained 13 additional positions, leading to a reporter with high-level PhoB-dependent expression, essentially no background, and no other known regulatory elements. As expressed reporters, we used both stable and destabilized variants of green fluorescent protein (GFP), the latter of which has a half-life of 81 min in V. fischeri . In culture, PhoB induced the reporter when phosphate was depleted to a concentration below 10 μM. During symbiotic colonization of its host squid, Euprymna scolopes , the reporter indicated heterogeneous phosphate availability in different light-organ microenvironments. Finally, testing this construct in other members of the Proteobacteria demonstrated its broader utility. The results illustrate how a limited ability to predict synthetic promoter-reporter performance can be overcome through iterative screening and reengineering. IMPORTANCE Transcriptional reporters can be powerful tools for assessing when a particular regulator is active; however, native promoters may not be ideal for this purpose. Optimal reporters should be specific to the regulator being examined and should maximize the difference between the “on” and “off” states; however, these properties are distinct from the selective pressures driving the evolution of natural promoters. Synthetic promoters offer a promising alternative, but our understanding often does not enable fully predictive promoter design, and the large number of alternative sequence possibilities can be intractable. In a synthetic promoter region with over 34 billion sequence variants, we identified bases correlated with favorable performance by screening only 83 candidates, allowing us to rationally constrain our design. We thereby generated an optimized reporter that is induced by PhoB and used it to explore the low-phosphate response of V. fischeri . This promoter design strategy will facilitate the engineering of other regulator-specific reporters.« less