An official website of the United States government
Surrogate based optimization (SBO) methods have gained popularity in the field of constrained optimization of expensive black-box functions. However, constraint handling methods do not usually guarantee strictly feasible candidates during optimization. This can become an issue in applied engineering problems where design variables must remain feasible for simulations to not fail. We propose a simple constraint-handling method for computationally inexpensive constraint functions which guarantees strictly feasible candidates when using a surrogate-based optimizer. We compare our method to other SBO algorithms and an EA on five analytical test functions, and an applied fully-resolved Computational Fluid Dynamics (CFD) problem concerned with optimization of an undulatory swimming of a fish-like body, and show that the proposed algorithm shows favorable results while guaranteeing feasible candidates.
more »
« less
Topological surrogates for computationally efficient seismic robustness optimization of water pipe networks
Abstract The criticality of seismic robustness of the water pipe networks cannot be overstated. Current methodologies for optimizing seismic robustness of city‐scale water pipe networks are scarce. A very few studies that can be found are also prone to long optimization runtimes due to the requirement of repeated hydraulic analysis. Hence, there is a critical need for the identification of computationally efficient surrogate optimization methods for maximizing seismic robustness of water pipe networks. To address this need, this research was conducted to identify, for the first time, computationally efficient topological surrogates for hydraulic simulation‐based optimization. The computational efficiency of surrogate optimization was measured in terms of solution quality (i.e., post‐earthquake serviceability) and computational runtime. Ten different topological connectivity metrics were evaluated out of which five were considered computationally infeasible due to their prohibitive optimization runtime. Five remaining metrics were then used to formulate five surrogate objective functions for seismic robustness of water pipe networks. Each of these functions was optimized using a simulated annealing‐based algorithm. Application of the proposed approach to city‐level benchmark networks helped to identify two metrics out of ten that offered a substantial reduction in optimization runtime with a minimal loss in solution quality. These findings will be highly valuable to water distribution network managers for identifying economical rehabilitation policies for enhancing the seismic robustness at a city‐scale within a reasonable amount of time.
more »
« less
- Award ID(s):
- 1926792
- PAR ID:
- 10157179
- Publisher / Repository:
- Elsevier
- Date Published:
- Journal Name:
- Computer-Aided Civil and Infrastructure Engineering
- Volume:
- 35
- Issue:
- 10
- ISSN:
- 1093-9687
- Format(s):
- Medium: X Size: p. 1101-1114
- Size(s):
- p. 1101-1114
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract We introduce the concept of decision‐focused surrogate modeling for solving computationally challenging nonlinear optimization problems in real‐time settings. The proposed data‐driven framework seeks to learn a simpler, for example, convex, surrogate optimization model that is trained to minimize thedecision prediction error, which is defined as the difference between the optimal solutions of the original and the surrogate optimization models. The learning problem, formulated as a bilevel program, can be viewed as a data‐driven inverse optimization problem to which we apply a decomposition‐based solution algorithm from previous work. We validate our framework through numerical experiments involving the optimization of common nonlinear chemical processes such as chemical reactors, heat exchanger networks, and material blending systems. We also present a detailed comparison of decision‐focused surrogate modeling with standard data‐driven surrogate modeling methods and demonstrate that our approach is significantly more data‐efficient while producing simple surrogate models with high decision prediction accuracy.more » « less
-
Maintaining the functionality of wastewater networks is critical to individual well-being, business continuity, public health, and safety. However, seismic damage and loss assessments of wastewater networks traditionally use fragility functions based on median repair rates without considering relevant sources of uncertainty and correlations of damage when estimating potential damage states and pipe repairs. This study presents a probabilistic methodology to incorporate modeling uncertainty (e.g. model parameter and model class uncertainty) and spatial correlations (e.g. spatial auto- and cross-correlation) of pipe repairs. The methodology was applied to a case study backbone system of a wastewater network in Portland, OR, using the expected hazard intensity maps for multiple deterministic earthquake scenarios, including a moment magnitude M6.8 Portland Hills Fault and M8.1, M8.4, M8.7, and M9.0 Cascadia Subduction Zone (CSZ) events. As spatial-correlation models of pipeline damage were non-existent in the literature and local information on costs to repair the pipes was limited at the time of this study, correlation methods and repair costs were proposed to estimate lower and upper bounds of pipe damage and loss. The results show how the consideration of different levels of uncertainty and spatial correlation for pipe repair rate could lead to different probabilistic estimates of damage and loss at the system level of the wastewater network, even though the point estimates, such as the mean and median, remain essentially unaltered.more » « less
-
null (Ed.)Abstract Cyber–physical–social systems (CPSS) with highly integrated functions of sensing, actuation, computation, and communication are becoming the mainstream consumer and commercial products. The performance of CPSS heavily relies on the information sharing between devices. Given the extensive data collection and sharing, security and privacy are of major concerns. Thus, one major challenge of designing those CPSS is how to incorporate the perception of trust in product and systems design. Recently, a trust quantification method was proposed to measure the trustworthiness of CPSS by quantitative metrics of ability, benevolence, and integrity. The CPSS network architecture can be optimized by choosing a subnet such that the trust metrics are maximized. The combinatorial network optimization problem, however, is computationally challenging. Most of the available global optimization algorithms for solving such problems are heuristic methods. In this paper, a surrogate-based discrete Bayesian optimization method is developed to perform network design, where the most trustworthy CPSS network with respect to a reference node is formed to collaborate and share information with. The applications of ability and benevolence metrics in design optimization of CPSS architecture are demonstrated.more » « less
-
Learning to optimize (L2O) has recently emerged as a promising approach to solving optimization problems by exploiting the strong prediction power of neural networks and offering lower runtime complexity than conventional solvers. While L2O has been applied to various problems, a crucial yet challenging class of problems — robust combinatorial optimization in the form of minimax optimization — have largely remained under-explored. In addition to the exponentially large decision space, a key challenge for robust combinatorial optimization lies in the inner optimization problem, which is typically non-convex and entangled with outer optimization. In this paper, we study robust combinatorial optimization and propose a novel learning-based optimizer, called LRCO (Learning for Robust Combinatorial Optimization), which quickly outputs a robust solution in the presence of uncertain context. LRCO leverages a pair of learning-based optimizers — one for the minimizer and the other for the maximizer — that use their respective objective functions as losses and can be trained without the need of labels for training problem instances. To evaluate the performance of LRCO, we perform simulations for the task offloading problem in vehicular edge computing. Our results highlight that LRCO can greatly reduce the worst-case cost and improve robustness, while having a very low runtime complexity.more » « less
