Stochastic Problem Research Articles

Stochastic golden ratio algorithm to non-convex stochastic mixed variational inequality problem

Stochastic golden ratio algorithm to non-convex stochastic mixed variational inequality problem

Adaptive Two-Stage Stochastic Programming with an Analysis on Capacity Expansion Planning Problem

Problem definition: Multistage stochastic programming is a well-established framework for sequential decision making under uncertainty by seeking policies that can be dynamically adjusted as uncertainty is realized. Often, for example, because of contractual constraints, such flexible policies are not desirable, and the decision maker may need to commit to a set of actions for a certain number of periods. Two-stage stochastic programming might be better suited to such settings, where first-stage decisions do not adapt to the uncertainty realized. In this paper, we propose a novel alternative approach, named as adaptive two-stage stochastic programming, where each component of the decision policy requiring limited flexibility has its own revision point, a period prior to which the decisions are determined at the beginning of the planning until this revision point, and after which they are revised for adjusting to the uncertainty realized thus far until the end of the planning. We then analyze this approach over the capacity expansion planning problem, that may require limited flexibility over expansion decisions. Methodology/results: We provide a generic mixed-integer programming formulation for the adaptive two-stage stochastic programming problem with finite support, in particular, for scenario trees, and show that this problem is NP-hard in general. Next, we focus on the capacity expansion planning problem and derive bounds on the value of adaptive two-stage programming in comparison with the two-stage and multistage approaches in terms of revision points. We propose several heuristic solution algorithms based on this bound analysis. These algorithms either provide approximation guarantees or computational advantages in solving the resulting adaptive two-stage stochastic problem. Managerial implications: We provide insights on the choice of the revision times based on our analytical analysis. We further present an extensive computational study on a generation capacity expansion planning problem with different generation resources including renewable energy. We demonstrate the value of adopting adaptive two-stage approach against the existing policies under limited flexibility and highlight the efficiency of the proposed heuristics along with practical implications on the studied problem. Funding: This work was supported by the National Science Foundation [Grant 1633196] and the Office of Naval Research [Grant N00014-18-1-2075]. Supplemental Material: The online appendices are available at https://doi.org/10.1287/msom.2023.0157 .

A Simple and Optimal Policy Design with Safety Against Heavy-Tailed Risk for Stochastic Bandits

We study the stochastic multi-armed bandit problem and design new policies that enjoy both optimal regret expectation and light-tailed risk for regret distribution. We first find that any policy that obtains the optimal instance-dependent expected regret could incur a heavy-tailed regret tail risk that decays slowly with T. We then focus on policies that achieve optimal worst-case expected regret. We design a novel policy that (i) enjoys the worst-case optimality for regret expectation and (ii) has the worst-case tail probability of incurring a regret larger than any regret threshold that decays exponentially with respect to T. The decaying rate is proved to be optimal for all worst-case optimal policies. Our proposed policy achieves a delicate balance between doing more exploration at the beginning of the time horizon and doing more exploitation when approaching the end, compared with standard confidence-bound-based policies. We also enhance the policy design to accommodate the “any-time” setting where T is unknown a priori, highlighting “lifelong exploration”, and prove equivalently desired policy performances as compared with the “fixed-time” setting with known T. From a managerial perspective, we show through numerical experiments that our new policy design yields similar efficiency and better safety compared to celebrated policies. Our policy design is preferable especially when (i) there is a risk of underestimating the volatility profile, or (ii) there is a challenge of tuning policy hyper-parameters. We conclude by extending our proposed policy design to the stochastic linear bandit setting that leads to both worst-case optimality in terms of regret expectation and light-tailed risk on regret distribution. This paper was accepted by J. George Shanthikumar, data science. Funding: The work of D. Simchi-Levi and F. Zhu is partially supported by the MIT Data Science Laboratory. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2022.03512 .

Inverse chance-constrained dynamic data envelopment analysis under natural and managerial disposability: Concerning renewable energy efficiency and potentials

The performance of dynamic systems with specific data has been examined in some data envelopment analysis (DEA) studies. However, in many real-life situations, dealing with dynamic systems becomes challenging because of random factors. So, in this paper, a chance-constrained dynamic DEA approach under managerial and natural disposability is first proposed to analyze the dynamic renewable energy efficacy of processes in the presence of random performance measures. Then, the potentials of some performance metrics are addressed for changes of others using the introduced inverse chance-constrained dynamic DEA models. The chance-constrained dynamic DEA techniques and their inverse dynamic stochastic problems are converted into linear problems. The introduced approaches are applied to probe the renewable energy efficiency and potentials of some OECD countries in a span of time. The results show stochastic dynamic DEA and inverse stochastic dynamic DEA provided are practical tools for making the best choices, when there is uncertainty.

Branch and Price for the Stochastic Traveling Salesman Problem with Generalized Latency

We consider an extension of the symmetric traveling salesman problem with generalized latency that explicitly models uncertainty. The stochastic traveling salesman problem with generalized latency (STSP-GL) aims to choose a subset of nodes of an undirected graph and determines a Hamiltonian tour among those nodes, minimizing an objective function that is a weighted combination of route design and passenger routing costs. These nodes are selected to ensure that a predefined percentage of uncertain passenger demand is served with a given probability. We formulate the STSP-GL as a stochastic program and propose a branch-and-price algorithm for solving its deterministic equivalent. We also develop a local search approach with which we improve the performance of the branch-and-price approach. We assess the efficiency of the proposed methods on a set of instances from the literature. We demonstrate that the proposed methods outperform a known benchmark, improving upper bounds by up to 85% and lower bounds by up to 55%. Finally, we show that solutions of the stochastic model are both more cost-effective and robust than those of the deterministic model. History: This paper has been accepted for the Transportation Science Special Issue on TSL Conference 2023. Funding: This work was supported by the Bundesministerium für Bildung und Forschung [Grant 03ZU1105FA]. Supplemental Material: The online appendix is available at https://doi.org/10.1287/trsc.2023.0417 .

Spatial-temporal deep learning method for solving data-driven multi-echelon stochastic lot sizing problems with intermediate demands

ABSTRACT For the production processes and inventory decisions of a multichannel supply chain, a multi-echelon stochastic lot sizing problem with intermediate demands is extended to a constrained problem that considers the production (inventory) capacity constraint and to multi-item problems that consider feature and item correlations. Three spatial-temporal deep learning methods combining convolutional, recurrent and graph convolutional neural networks are proposed to transform the extended problems as spatial-temporal sequential prediction problems and learn data-driven decisions. Real-world data are used in the computational experiments, and the results show that considering both temporal and spatial features and/or feature correlations can provide improved prediction performance.

Invariant cones for jump-diffusions in infinite dimensions

In this paper we provide sufficient conditions for stochastic invariance of closed convex cones for stochastic partial differential equations (SPDEs) of jump-diffusion type, and clarify when these conditions are necessary. We emphasize that the diffusion coefficient of the SPDE does not need to be smooth, an assumption which is frequently imposed when dealing with stochastic invariance problems. Our results apply to the positive cone of abstract L2\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$L^2$$\\end{document}-spaces. Furthermore, we present a series of applications, where we investigate SPDEs arising in natural sciences and economics.

Calibration of stochastic, agent-based neuron growth models with approximate Bayesian computation.

Understanding how genetically encoded rules drive and guide complex neuronal growth processes is essential to comprehending the brain's architecture, and agent-based models (ABMs) offer a powerful simulation approach to further develop this understanding. However, accurately calibrating these models remains a challenge. Here, we present a novel application of Approximate Bayesian Computation (ABC) to address this issue. ABMs are based on parametrized stochastic rules that describe the time evolution of small components-the so-called agents-discretizing the system, leading to stochastic simulations that require appropriate treatment. Mathematically, the calibration defines a stochastic inverse problem. We propose to address it in a Bayesian setting using ABC. We facilitate the repeated comparison between data and simulations by quantifying the morphological information of single neurons with so-called morphometrics and resort to statistical distances to measure discrepancies between populations thereof. We conduct experiments on synthetic as well as experimental data. We find that ABC utilizing Sequential Monte Carlo sampling and the Wasserstein distance finds accurate posterior parameter distributions for representative ABMs. We further demonstrate that these ABMs capture specific features of pyramidal cells of the hippocampus (CA1). Overall, this work establishes a robust framework for calibrating agent-based neuronal growth models and opens the door for future investigations using Bayesian techniques for model building, verification, and adequacy assessment.

Leader-Following Sampled-Data Consensus of Multiagent Systems With Successive Packet Losses and Stochastic Sampling.

In practical applications, sampled-data systems are often affected by unforeseen physical constraints that cause the sampling interval to deviate from the expected value and fluctuate according to a certain probability distribution. This probability distribution can be determined in advance through statistical analysis. Taking into account this stochastic sampling interval, this article focuses on addressing the leader-following sampled-data consensus problem for linear multiagent systems (MASs) with successive packet losses. First, the relationship of the equivalent sampling interval between two successive update instants is established, taking into account the double randomness introduced by both SPLs and stochastic sampling. Then, the equivalent discrete-time MAS is obtained, and the overall leader-following consensus problem is formulated as a stochastic stability problem of the equivalent system by incorporating the sampled-data consensus protocol and properties of the Laplacian matrix. Based on the equivalent the discrete-time system, a consensus criterion is derived under a directed graph by using the Lyapunov theory and leveraging a vectorization technique. By the introduction of a matrix reconstruction approach, the mathematical expectation of a product of three matrices, including the system matrix and its transpose, can be determined. Then, the consensus protocol gain is designed. Finally, an example is provided to validate our theoretical results.

Free vibrations of a Continuous-discrete multi-span Beam taking into account inertial Rotation Forces

Objective. The aim of the study is to estimate free vibrations of a continuousdiscrete multi-span beam taking into account the inertial forces of rotation. The goal is to determine the spectra of natural frequencies, damping coefficients and natural modes. Method. The study is based on the methods of linear mechanics of structures; numerical and numericalanalytical calculation methods. The solution to the problem is found using the method of separation of variables. Rotational movements of particles of continuous sections are taken into account according to one of the models of the Timoshenko beam. The D'Alembert principle and hypotheses on the smallness of displacements and angles of rotation of sections are used. Result. A system of equations in matrix-vector form is obtained. The mathematical model of transverse vibrations consists of three systems of differential equations. The equation includes transverse forces, external concentrated forces, d'Alembert inertial forcesi, and linear-viscous resistance forces. The inertial forces of rotation of concentrated masses are taken into account. The boundary and other additional conditions to the equations correspond to the calculation scheme. The left end of the beam is hinged. The conjugation conditions are met at the junction of the sections. Conclusions. This random process of disturbances is very close to the processes used in deterministic problems. The amplitudes and standard deviations of displacements in the deterministic and stochastic problems almost coincide, which confirms the reliability of the proposed calculation theory. Analysis of the curves shows that the standard deviations significantly depend on the degree of correlation of the components of the vector random process of disturbances. The use of modern computing computer systems such as Matlab allows us to successfully combine the advantages of both numerical and graphical methods.

Markov chain Monte Carlo methods applied to the stochastic inversion of 1D viscoelastic parameters

Abstract The characterization of the subsurface profile properties that are susceptible to liquefaction during earthquakes is of great importance in accident prevention. Field data can be used in a stochastic inverse problem to estimate the material properties using Markov chain Monte Carlo Methods (McMC) that incorporates prior knowledge about the unknown parameters as well as the data available. Here, we numerically study the Random Walk (RW) and Differential Evolution (DE) variations of the Metropolis algorithm in the context of seismic modeling considering the viscoelastic wave equation. The algorithms are tested with data from the 1987 Superstition Hills earthquake recorded at the Wildlife Site. The results show the effectiveness and accuracy of these algorithms, with emphasis to DE which reached convergence earlier compared to RW.

Stochastic dual dynamic programming for optimal power flow problems under uncertainty

Planning in the power sector has to take into account the physical laws of alternating current (AC) power flows as well as uncertainty in the data of the problems, both of which greatly complicate optimal decision making. We propose a computationally tractable framework to solve multi-stage stochastic optimal power flow (OPF) problems in AC power systems. Our approach uses recent results on dual convex semi-definite programming (SDP) relaxations of OPF problems in order to adapt the stochastic dual dynamic programming (SDDP) algorithm for problems with a Markovian structure. We show that the usual SDDP lower bound remains valid and that the algorithm converges to a globally optimal policy of the stochastic AC-OPF problem as long as the SDP relaxations are tight. To test the practical viability of our approach, we set up a case study of a storage siting, sizing, and operations problem. We show that the convex SDP relaxation of the stochastic problem is usually tight and discuss ways to obtain near-optimal physically feasible solutions when this is not the case. The algorithm finds a physically feasible policy with an optimality gap of 3% and yields a significant added value of 27% over a rolling deterministic policy, which leads to overly optimistic policies and underinvestment in flexibility. This suggests that the common industry practice of assuming direct current and deterministic problems should be reevaluated by considering models that incorporate realistic AC flows and stochastic elements in the data as potentially more realistic alternatives.

Optimal solution of multiobjective stable environmental economic power dispatch problem considering probabilistic wind and solar PV generation

The Environmental Economic Power Dispatch (EEPD) problem, a widely studied bi-objective nonlinear optimization challenge in power systems, traditionally focuses on the economic dispatch of thermal generators without considering network security constraints. However, environmental sustainability necessitates reducing emissions and increasing the penetration of renewable energy sources (RES) into the electrical grid. The integration of high levels of RES, such as wind and solar PV, introduces stability issues due to their uncertain and intermittent nature. This paper addresses these concerns by formulating and solving the Stable Environmental Economic Power Dispatch (SEEPD) problem, which includes fixed zonal reserve capacity from conventional thermal generators and uncertain reserves from RES. Uncertainties in RES and load demand are modeled using random variable generation techniques, applying Gaussian, Weibull, and log-normal probability density functions (PDFs) for load demand, wind velocity, and solar irradiance, respectively. The stochastic SEEPD problem extends to multiple periods by replicating the single-period problem for each interval in the planning horizon, linking periods through intertemporal ramping costs, physical ramp rate, and fixed zonal reserve constraints on dispatch variables. Multi-Objective Evolutionary Algorithms (MOEAs) have gained importance for solving complex nonlinear problems involving multi-objective functions. This paper applies the latest MOEAs to tackle the proposed SEEPD problem, incorporating stochastic wind and solar PV power sources. Network security constraints, such as transmission line capacities and bus voltage limits, are considered along with constraints on generator capabilities and intertemporal spinning reserves, ramp-up and ramp-down constraints for thermal generators. A bidirectional coevolutionary-based multi-objective evolutionary algorithm is employed, integrating an advanced constraint-handling technique to ensure compliance with system constraints. The simulation results show that the proposed formulation achieves a better trade-off between various conflicting objective functions compared to other state-of-the-art MOEAs.

An Oscillation-Free Discontinuous Galerkin Method for a Nonlinear Stochastic Convection-Dominated Diffusion Problem and Its Error Analysis

An Oscillation-Free Discontinuous Galerkin Method for a Nonlinear Stochastic Convection-Dominated Diffusion Problem and Its Error Analysis

Policy with guaranteed risk-adjusted performance for multistage stochastic linear problems

Policy with guaranteed risk-adjusted performance for multistage stochastic linear problems

Lossy Image Compression with Stochastic Quantization

Introduction. Lossy image compression algorithms play a crucial role in various domains, including graphics, and image processing. As image information density increases, so do the resources required for processing and transmission. One of the most prominent approaches to address this challenge is color quantization, proposed by Orchard et al. (1991). This technique optimally maps each pixel of an image to a color from a limited palette, maintaining image resolution while significantly reducing information content. Color quantization can be interpreted as a clustering problem (Krishna et al. (1997), Wan (2019)), where image pixels are represented in a three-dimensional space, with each axis corresponding to the intensity of an RGB channel. The purpose of the paper. Scaling of traditional algorithms like K-Means can be challenging for large data, such as modern images with millions of colors. This paper reframes color quantization as a three-dimensional stochastic transportation problem between the set of image pixels and an optimal color palette, where the number of colors is a predefined hyperparameter. We employ Stochastic Quantization (SQ) with a seeding technique proposed by Arthur et al. (2007) to enhance the scalability of color quantization. This method introduces a probabilistic element to the quantization process, potentially improving efficiency and adaptability to diverse image characteristics. Results. To demonstrate the efficiency of our approach, we present experimental results using images from the ImageNet dataset. These experiments illustrate the performance of our Stochastic Quantization method in terms of compression quality, computational efficiency, and scalability compared to traditional color quantization techniques. Conclusions. This study introduces a scalable algorithm for solving the color quantization problem without memory constraints, demonstrating its efficiency on a subset of images from the ImageNet dataset. The convergence speed of the algorithm can be further enhanced by modifying the update rule with alternative methods to Stochastic Gradient Descent (SGD) that incorporate adaptive learning rates. Moreover, the stochastic nature of the proposed solution enables the utilization of parallelization techniques to simultaneously update the positions of multiple quants, potentially leading to significant performance improvements. This aspect of parallelization and its impact on algorithm efficiency presents a topic for future research. The proposed method not only addresses the limitations of existing color quantization techniques but also opens up new possibilities for optimizing image compression algorithms in resource-constrained environments. Keywords: non-convex optimization, stochastic optimization, stochastic quantization, color quantization, lossy compression.

Stochastic Resource Allocation Problems: Minmax and Maxmin Solutions

Problem definition: Min-max or max-min objective criteria naturally arise in many resource allocation applications, especially when the attendant goal involves distribution of a limited resource across multiple subsystems, locations, or activities. Each has a different consumption rate of the resource. In this setting, the decision maker is interested in optimizing the system performance, which can be measured in terms of the smallest or largest order statistic across subsystems. Examples include the allocation of health resources to maximize the first runout time or to minimize the maximum delay. Methodology/results: In the literature, there are no unified theoretical and computational results on the optimal allocations for these problems. Thus, in this paper, we develop key theoretical properties of the attendant stochastic resource allocation problems, providing a technical condition for the objective functions to have favorable properties, upon which we derive novel optimal allocation rules with clear interpretation. Managerial implications: The proportional allocation rule can be interpreted as a robust solution when the full distribution of the interarrival time is unknown. It generally has poor performance, especially when the resource consumption rates vary across locations, which happens regularly in practice. Our allocation rule is exact and should be used for the allocation of scarce resources—that is, when the number of resources per location is limited, which is often encountered in healthcare settings.

Dynamic Programming, Neuro-Dynamic Programming, Rollout Method and Model Predictive Control to Optimal Control of a Fermentation Process

This work presents the use of dynamic programming (DP), neuro-dynamic programming (NDP), rollout algorithm (RA) and model predictive control (MPC) for optimal control of batch cultivation of the yeast Kluyveromyces marxianus var. lactis MC5. DP is a widespread method for solving problems related to optimization and optimal process control. To reduce the “curse of dimensionality”, NDP has been implemented as an alternative. In NDP, a neural network is used to solve the dimensionality problem. A simpler NDP method, called RA, is used to approximate the optimal cost through the cost of a relatively good suboptimal policy, called the baseline policy. RA is a suboptimal method for deterministic and stochastic problems that can be solved by DP. In this paper we also present off-line MPC technique for tracking of constrained fermentation systems and it overcomes the problem by off-line optimizations prior to implementation. MPC is used to provide perturbation feedback and it is developed theoretical on base a controller as an illustration how we can avoid disturbances in the process optimisation. The developed control algorithm-combined NDP and MPC ensures maximum biomass production at the end of the process and feedback during disturbances and process stability and shows that robust stability can be ensured.

A new multi-objective-stochastic framework for reconfiguration and wind energy resource allocation in distribution network incorporating improved dandelion optimizer and uncertainty

Improving the reliability and power quality of unbalanced distribution networks is crucial for ensuring consistent and reliable electricity supply. In this research, multi-objective optimization of unbalanced distribution networks reconfiguration integrated with wind turbine allocation (MORWTA) is implemented considering uncertainties of networks load, and also wind power incorporating a stochastic framework. The multi-objective function is defined by the minimization of power loss, voltage sag (VS), total harmonic distortion (THD), voltage unbalance (VU), energy not-supplied (ENS), system average interruption frequency index (SAIFI), system average interruption duration index (SAIDI), and momentary average interruption frequency (MAIFI). A new improved dandelion optimizer (IDO) with adaptive inertia weight is recommended to counteract premature convergence to identify decision variables, including the optimal network configuration through opened switches and the best location and size of wind turbines in the networks. The stochastic problem is modeled using the 2m + 1 point estimate method (PEM) combined with K-means clustering, taking into account the mentioned uncertainties. The proposed stochastic methodology is implemented on three modified 33-bus, and unbalanced 25-, and 37-bus distribution networks. The results demonstrated that the MORWTA enhanced all study objectives in comparison to the base networks. The results also demonstrated that the IDO had superior capability to solve the deterministic- and stochastic-MORWTA in comparison to the conventional DO, grey wolf optimizer (GWO), particle swarm optimization (PSO), and arithmetic optimization algorithm (AOA) in terms of achieving greater objective value. Moreover, the results demonstrated that when the stochastic-MORWTA model is considered, the power loss, VS, THD, VU, ENS, SAIFI, SAIDI, and MAIFI are increased by 18.35%, 9.07%, 10.43%, 12.46%, 11.90%, 9.28%, 12.16% and 14.36%, respectively for 25-bus network, and also these objectives are increased by 12.21%, 10.64%, 12.37%, 9.82%, 14.30%, 12.65%, 12.63% and 13.89%, respectively for 37-bus network compared to the deterministic-MORWTA model, which is related to the defined uncertainty patterns.

Fleet repositioning in the tramp ship routing and scheduling problem with bunker optimization: A matheuristic solution approach

This paper investigates an important planning problem faced by dry bulk shipping operators, referred to as the Tramp Ship Routing and Scheduling Problem with Bunker Optimization (TSRSPBO). The problem is to maximize the overall profit of a fleet of vessels by selecting cargoes and determining ship routes and schedules. We consider this problem under a set of practically relevant features such as flexibility in cargo quantities, as well as bunkering decisions on where to procure fuel and how much. As a particularly novel feature, we address the regional allocation of vessels at the end of the planning period to be well prepared for meeting (uncertain) future demand. To incorporate this, we consider the TSRSPBO as a two-stage stochastic programming problem, where cargo selection, routing, and bunkering decisions are solved in the first-stage problem, and the recourse cost of fleet repositioning is considered in the second stage. We present arc flow and path flow formulations, where the latter employs a priori generation of feasible routes as input. For solving realistically sized instances, we propose a matheuristic based on an Adaptive Large Neighborhood Search (ALNS) framework that iteratively generates columns and solves the path flow model. Computational experiments based on real data show that this matheuristic finds high-quality solutions for large test instances with 120 cargoes, 30 vessels, and ten bunker ports in less than one hour. Also, considering the TSRSPBO as a two-stage stochastic problem achieves the highest profits and is solved almost as quickly as the deterministic problem variant.