Posts by Collection

portfolio

publications

High-Order Rogue Waves and Their Dynamics of the Fokas–Lenells Equation Revisited: a Variable Separation Technique

Published in Nonlinear Dynamics, 2019

Abstract

The Fokas–Lenells (FL) equation is an integrable higher-order extension of nonlinear Schrödinger equation. One approach to generating its breather solutions is based on Darboux transformation (DT) and iterations. However, the DT of FL equation contains negative powers of the spectral parameter, which can lead to very complicated expressions when $N$ is large. In this paper, we avoid the negative powers by adopting a variable separation and Taylor expansion technique to solve the Lax pair of FL system. Furthermore, stability of the proposed technique is demonstrated in detail.

Download Paper

Nonparametric Kullback-Liebler Divergence Estimation Using M-Spacing

Published in 2021 Winter Simulation Conference, 2021

Abstract

Entropy of a random variable with unknown distribution function can be estimated nonparametrically by spacing methods when independent and identically distributed (i.i.d.) observations of the random variable are available. We extend the classical entropy estimator based on sample spacing to define an m-spacing estimator for the Kullback-Liebler (KL) divergence between two i.i.d. observations with unknown distribution functions, which can be applied to measure discrepancy between real-world system output and simulation output as well as between two simulators’ outputs. We show that the proposed estimator converges almost surely to the true KL divergence as the numbers of outputs collected from both systems increase under mild conditions and discuss the required choices for $m$ and the simulation output sample size as functions of the real-world sample size. Additionally, we show Central Limit Theorems for the proposed estimator with appropriate scaling.

Download Paper

Nonparametric Learning of Two-Layer ReLU Residual Units

Published in Transactions on Machine Learning Research, 2022

Abstract

We describe an algorithm that learns two-layer residual units using rectified linear unit (ReLU) activation: suppose the input $\mathbf{x}$ is from a distribution with support space $\mathbb{R}^d$ and the groundtruth generative model is a residual unit of this type, given by $\mathbf{y}=\mathbf{B}^∗[(\mathbf{A}^∗{\mathbf{x}})^+ + \mathbf{x}]$, where ground-truth network parameters $\mathbf{A}^* \in \mathbb{R}^{d\times d}$ represent a full-rank matrix with nonnegative entries and $\mathbf{B}^* \in\mathbb{R}^{m\times d}$ is full-rank with $m \geq d$ and for $\mathbf{c} \in \mathbb{R}^d$, $[\mathbf{c}^+]_i = \max{0,c_i}$. We design layer-wise objectives as functionals whose analytic minimizers express the exact ground-truth network in terms of its parameters and nonlinearities. Following this objective landscape, learning residual units from finite samples can be formulated using convex optimization of a nonparametric function: for each layer, we first formulate the corresponding empirical risk minimization (ERM) as a positive semi-definite quadratic program (QP), then we show the solution space of the QP can be equivalently determined by a set of linear inequalities, which can then be efficiently solved by linear programming (LP). We further prove the strong statistical consistency of our algorithm, and demonstrate its robustness and sample efficiency through experimental results on synthetic data and a set of benchmark regression datasets.

Download Paper

Efficient Input Uncertainty Quantification for Regenerative Simulation

Published in 2023 Winter Simulation Conference, 2023

Abstract

The initial bias in steady-state simulation can be characterized as the bias of a ratio estimator if the simulation model has a regenerative structure. This work tackles input uncertainty quantification for a regenerative simulation model when its input distributions are estimated from finite data. Our aim is to construct a bootstrap-based confidence interval (CI) for the true simulation output mean performance that provides a correct coverage with significantly less computational cost than the traditional methods. Exploiting the regenerative structure, we propose a $k$-nearest neighbor ($k$NN) ratio estimator for the steady-state performance measure at each set of bootstrapped input models and construct a bootstrap CI from the computed estimators. Asymptotically optimal choices for $k$ and bootstrap sample size are discussed. We further improve the CI by combining the $k$NN and likelihood ratio methods. We empirically compare the efficiency of the proposed estimators with the standard estimator using queueing examples.

Download Paper

Stochastic Approximation for Multi-period Simulation Optimization with Streaming Input Data

Published in ACM Transactions on Modeling and Computer Simulation, 2024

Abstract

We consider a continuous-valued simulation optimization (SO) problem, where a simulator is built to optimize an expected performance measure of a real-world system while parameters of the simulator are estimated from streaming data collected periodically from the system. At each period, a new batch of data is combined with the cumulative data and the parameters are re-estimated with higher precision. The system requires the decision variable to be selected in all periods. Therefore, it is sensible for the decision-maker to update the decision variable at each period by solving a more precise SO problem with the updated parameter estimate to reduce the performance loss with respect to the target system. We define this decision-making process as the multi-period SO problem and introduce a multi-period stochastic approximation (SA) framework that generates a sequence of solutions. Two algorithms are proposed: Re-start SA (ReSA) reinitializes the stepsize sequence in each period, whereas Warm-start SA (WaSA) carefully tunes the stepsizes, taking both fewer and shorter gradient-descent steps in later periods as parameter estimates become increasingly more precise. We show that under suitable strong convexity and regularity conditions, ReSA and WaSA achieve the best possible convergence rate in expected sub-optimality either when an unbiased or a simultaneous perturbation gradient estimator is employed, while WaSA accrues significantly lower computational cost as the number of periods increases. In addition, we present the regularized ReSA which obviates the need to know the strong convexity constant and achieves the same convergence rate at the expense of additional computation.

Download Paper

Introductory Tutorial: Simulation Optimization under Input Uncertainty

Published in 2024 Winter Simulation Conference, 2024

Abstract

Input uncertainty in the simulation output is caused by the estimation error in the input models of the simulator due to finiteness of the data from which they are estimated. Ignoring input uncertainty when formulating and solving a simulation optimization problem may lead to a solution with poor system performance. This tutorial discusses how to incorporate input uncertainty in simulation optimization to avoid such risk. We first categorize the problems into three groups based on their contexts: fixed batch data, streaming data, and active input data collection problems. Input and simulation output response modeling frameworks that can be adopted in all three categories are discussed. Then, we provide a high-level overview of simulation optimization problem formulations and algorithmic approaches to tackle problems in each group. Some thoughts on future research directions are shared.

Download Paper

Digital Twin Validation with Multi-Epoch, Multi-Variate Output Data

Published in 2024 Winter Simulation Conference, 2024

Abstract

This paper studies validation of a simulation-based process digital twin (DT). We assume that at any point the DT is queried, the system state is recorded. Then, the DT simulator is initialized to match the system state and the simulations are run to predict the key performance indicators (KPIs) at the end of each time epoch of interest. Our validation question is if the distribution of the simulated KPIs matches that of the system KPIs at every epoch. Typically, these KPIs are multi-variate random vectors and non-identically distributed across epochs making it difficult to apply the existing validation methods. We devise a hypothesis test that compares the marginal and joint distributions of the KPI vectors, separately, by transforming the multi-epoch data to identically distributed observations. We empirically demonstrate that the test has good power when the system and the simulator sufficiently differ in distribution.

Download Paper

Efficient Input Uncertainty Quantification for Ratio Estimator

Published in INFORMS Journal on Computing, 2026

Abstract

We study the construction of a confidence interval (CI) for a simulation output performance measure that accounts for input uncertainty when the input models are estimated from finite data. In particular, we focus on performance measures that can be expressed as a ratio of two dependent simulation outputs’ means. We adopt the parametric bootstrap method to mimic input data sampling and construct the percentile bootstrap CI after estimating the ratio at each bootstrap sample. The standard estimator, which takes the ratio of two sample means, tends to exhibit large finite-sample bias and variance, leading to overcoverage of the percentile bootstrap CI. To address this, we propose two new ratio estimators that replace the sample means with pooled mean estimators via the k-nearest neighbor (kNN) regression: the kNN estimator and the kLR estimator. The kNN estimator performs well in low dimensions, but its estimation error converges more slowly as the dimension increases. The kLR estimator combines the likelihood ratio (LR) method with the kNN regression, leveraging the strengths of both while mitigating their weaknesses; the LR method removes dependence of the error convergence rate on the dimension, whereas the kNN method controls the variance of the kLR estimator to be asymptotically bounded. From the asymptotic analyses and finite-sample heuristics, we propose an experiment design for the ratio estimators and demonstrate their superior empirical performances over the standard ratio estimator using three examples, including one in the enterprise risk management application.

Download Paper

talks

teaching

Teaching experience 1

Undergraduate course, University 1, Department, 2014

This is a description of a teaching experience. You can use markdown like any other post.

Teaching experience 2

Workshop, University 1, Department, 2015

This is a description of a teaching experience. You can use markdown like any other post.