arXiv subjects

David A. Goldberg

Publications and source records attributed to David A. Goldberg.

15 recordsLinked to original sources

arXiv2026

Online Stochastic Packing and Combinatorial Allocation with General Correlations

There has been a growing interest in studying online stochastic packing and combinatorial allocation under more general correlation structures, motivated by the complex data sets and models driving modern applications. Several past works either assume correlations are weak or have a particular structure, have a complexity scaling with the number of Markovian ``states of the world" (which may be exponentially large in the case of full history dependence), scale poorly with the horizon $T$, or make additional continuity assumptions. We show that for all $ε$, the online stochastic packing (combinatorial allocation) linear programming problem with general correlations (suitably normalized and with sparse columns) has an approximately optimal policy (with optimality gap $εT$) whose per-decision runtime scales as the time to simulate a single sample path of the underlying stochastic process (assuming access to a Monte Carlo simulator), multiplied by a constant independent of the horizon or number of Markovian states. We derive analogous results for network revenue management (also with flexible products), and online bipartite matching and independent set in bounded-degree graphs, by rounding. Our algorithms implement stochastic gradient methods in a novel on-the-fly/recursive manner for the associated massive deterministic-equivalent linear program.

math.OC↗

arXiv2024

Polynomial time algorithm for optimal stopping with fixed accuracy

The problem of high-dimensional path-dependent optimal stopping (OS) is important to multiple academic communities and applications. Modern OS tasks often have a large number of decision epochs, and complicated non-Markovian dynamics, making them especially challenging. Standard approaches, often relying on ADP, duality, deep learning and other heuristics, have shown strong empirical performance, yet have limited rigorous guarantees (which may scale exponentially in the problem parameters and/or require previous knowledge of basis functions or additional continuity assumptions). Although past work has placed these problems in the framework of computational complexity and polynomial-time approximability, those analyses were limited to simple one-dimensional problems. For long-horizon complex OS problems, is a polynomial time solution even theoretically possible? We prove that given access to an efficient simulator of the underlying information process, and fixed accuracy epsilon, there exists an algorithm that returns an epsilon-optimal solution (both stopping policies and approximate optimal values) with computational complexity scaling polynomially in the time horizon and underlying dimension. Like the first polynomial-time (approximation) algorithms for several other well-studied problems, our theoretical guarantees are polynomial yet impractical. Our approach is based on a novel expansion for the optimal value which may be of independent interest.

math.PR↗

arXiv2023

Simple and explicit bounds for multi-server queues with $1/(1-ρ)$ scaling

We consider the FCFS $GI/GI/n$ queue, and prove the first simple and explicit bounds that scale as $\frac{1}{1-ρ}$ under only the assumption that inter-arrival times have finite second moment, and service times have finite $2+ε$ moment for some $ε> 0$. Here $ρ$ denotes the corresponding traffic intensity. Conceptually, our results can be viewed as a multi-server analogue of Kingman's bound. Our main results are bounds for the tail of the steady-state queue length and the steady-state probability of delay. The strength of our bounds (e.g. in the form of tail decay rate) is a function of how many moments of the service distribution are assumed finite. Our bounds scale gracefully even when the number of servers grows large and the traffic intensity converges to unity simultaneously, as in the Halfin-Whitt scaling regime. Some of our bounds scale better than $\frac{1}{1-ρ}$ in certain asymptotic regimes. In these same asymptotic regimes we also prove bounds for the tail of the steady-state number in service. Our main proofs proceed by explicitly analyzing the bounding process which arises in the stochastic comparison bounds of Gamarnik and Goldberg for multi-server queues. Along the way we derive several novel results for suprema of random walks and pooled renewal processes which may be of independent interest. We also prove several additional bounds using drift arguments (which have much smaller pre-factors), and point out a conjecture which would imply further related bounds and generalizations. We also show that when all moments of the service distribution are finite and satisfy a mild growth rate assumption, our bounds can be strengthened to yield explicit tail estimates decaying as $O\big(\exp(-x^α)\big)$, with $α\in (0,1)$ depending on the growth rate of these moments.

math.PR↗

arXiv2018

Time (in)consistency of multistage distributionally robust inventory models with moment constraints

Recently, there has been a growing interest in developing inventory control policies which are robust to model misspecification. One approach is to posit that nature selects a worst-case distribution for any stochastic primitives from some pre-specified family. Several communities have observed that a subtle phenomena known as time inconsistency can arise in this framework. In particular, it becomes possible that a policy which is optimal at time zero may not be optimal for the associated optimization problem in which the decision-maker recomputes her policy at each point in time, which has implications for implementability. If there exists a policy which is optimal for both formulations, we say that the policy is time consistent, and the problem is weakly time consistent. If every optimal policy is time consistent, we say that the problem is strongly time consistent. We study these phenomena in the context of managing an inventory over time, when only the mean, variance, and support are known for the demand at each stage. We provide several illustrative examples showing that here the question of time consistency can be quite subtle, and complement these observations by providing simple sufficient conditions for weak and strong time consistency. Interestingly, our results show that time consistency may hold even when rectangularity does not. Although a similar phenomena was previously identified by Shapiro for the setting in which only the mean and support of the demand are known, there the problem was always weakly time consistent, with both formulations having the same optimal value. Here our model is rich enough to exhibit a variety of interesting behaviors, including lack of weak time consistency, strong time consistency even when both formulations have different optimal values, and non-existence of even a single optimal base-stock policy under the static formulation.

math.OC↗

arXiv2018

Distributionally robust inventory control when demand is a martingale

Demand forecasting plays an important role in many inventory control problems. To mitigate the potential harms of model misspecification, various forms of distributionally robust optimization have been applied. Although many of these methodologies suffer from the problem of time-inconsistency, the work of Klabjan et al. established a general time-consistent framework for such problems by connecting to the literature on robust Markov decision processes. Motivated by the fact that many forecasting models exhibit special structure, as well as a desire to understand the impact of positing different dependency structures, in this paper we formulate and solve a time-consistent distributionally robust multi-stage newsvendor model which naturally unifies and robustifies several inventory models with forecasting. In particular, many simple models of demand forecasting have the feature that demand evolves as a martingale. We consider a robust variant of such models, in which the sequence of future demands may be any martingale with given mean and support. Under such a model, past realizations of demand are naturally incorporated into the structure of the uncertainty set going forwards. We explicitly compute the minimax optimal policy (and worst-case distribution) in closed form, by combining ideas from convexity, probability, and dynamic programming. We prove that at optimality the worst-case demand distribution corresponds to the setting in which inventory may become obsolete, a scenario of practical interest. To gain further insight, we prove weak convergence (as the time horizon grows large) to a simple and intuitive process. We also compare to the analogous setting in which demand is independent across periods (analyzed previously by Shapiro), and identify interesting differences between these models, in the spirit of the price of correlations studied by Agrawal et al.

math.PR↗

arXiv2018

Large deviations analysis for the $M/H_2/n + M$ queue in the Halfin-Whitt regime

We consider the FCFS $M/H_2/n + M$ queue in the Halfin-Whitt heavy traffic regime. It is known that the normalized sequence of steady-state queue length distributions is tight and converges weakly to a limiting random variable W. However, those works only describe W implicitly as the invariant measure of a complicated diffusion. Although it was proven by Gamarnik and Stolyar that the tail of W is sub-Gaussian, the actual value of $\lim_{x \rightarrow \infty}x^{-2}\log(P(W >x))$ was left open. In subsequent work, Dai and He conjectured an explicit form for this exponent, which was insensitive to the higher moments of the service distribution. We explicitly compute the true large deviations exponent for W when the abandonment rate is less than the minimum service rate, the first such result for non-Markovian queues with abandonments. Interestingly, our results resolve the conjecture of Dai and He in the negative. Our main approach is to extend the stochastic comparison framework of Gamarnik and Goldberg to the setting of abandonments, requiring several novel and non-trivial contributions. Our approach sheds light on several novel ways to think about multi-server queues with abandonments in the Halfin-Whitt regime, which should hold in considerable generality and provide new tools for analyzing these systems.

math.PR↗

arXiv2017

Heavy-tailed queues in the Halfin-Whitt regime

We consider the FCFS G/G/n queue in the Halfin-Whitt regime, in the presence of heavy-tailed distributions (i.e. infinite variance). We prove that under minimal assumptions, i.e. only that processing times have finite 1 + epsilon moment and inter-arrival times have finite second moment, the sequence of stationary queue length distributions, normalized by $n^{\frac{1}{2}}$, is tight. All previous tightness results for the stationary queue length required that processing times have finite 2 + epsilon moment. Furthermore, we develop simple and explicit bounds on the stationary queue length in that setting. When processing times have an asymptotically Pareto tail with index alpha in (1,2), we bound the large deviations behavior of the limiting process, and derive a matching lower bound when inter-arrival times are Markovian. Interestingly, we find that the large deviations behavior of the limit has a sub-exponential decay, differing fundamentally from the exponentially decaying tails known to hold in the light-tailed setting, and answering an open question of Gamarnik and Goldberg. For the setting where instead the inter-arrival times have an asymptotically Pareto tail with index alpha in (1,2), we extend recent results of Hurvich and Reed (who analyzed the case of deterministic processing times) by proving that for general processing time distributions, the sequence of stationary queue length distributions, normalized by $n^{\frac{1}α}$, is tight (here we use the scaling of Hurvich and Reed, i.e. Halfin-Whitt-Reed regime). We are again able to bound the large-deviations behavior of the limit, and find that our derived bounds do not depend on the particular processing time distribution, and are in fact tight even for the case of deterministic processing times.

math.PR↗

arXiv2016

On the steady-state probability of delay and large negative deviations for the $GI/GI/n$ queue in the Halfin-Whitt regime

We consider the FCFS $GI/GI/n$ queue in the Halfin-Whitt heavy traffic regime, and prove bounds for the steady-state probability of delay (s.s.p.d.) for generally distributed processing times. We prove that there exist $ε_1, ε_2 > 0$, depending on the first three moments of the inter-arrival and processing time distributions, such that the s.s.p.d. is bounded from above by $\exp\big(-ε_1 B^2\big)$ as the associated excess parameter $B \rightarrow \infty$; and by $1 - ε_2 B$ as $B \rightarrow 0$. We also prove that the tail of the steady-state number of idle servers has a Gaussian decay. We provide explicit bounds in all cases, in terms of the first three moments of the inter-arrival and service distributions, and use known results to show that our bounds correctly capture various qualitative scalings. \\\indent Our main proof technique is the derivation of new stochastic comparison bounds for the FCFS $GI/GI/n$ queue, which are of a structural nature, hold for all $n$ and times $t$, and significantly generalize the work of \citet{GG.10c} (e.g. by providing bounds for the queue length to exceed any given level, as opposed to any given level strictly greater than the number of servers as acheived in \citet{GG.10c}). Our results do not follow from simple comparison arguments to e.g. infinite-server systems or loss models, which would in all cases provide bounds in the opposite direction.

math.PR↗

arXiv2016

Optimality gap of constant-order policies decays exponentially in the lead time for lost sales models

Inventory models with lost sales and large lead times have traditionally been considered intractable due to the curse of dimensionality. Recently, Goldberg and co-authors laid the foundations for a new approach to solving these models, by proving that as the lead time grows large, a simple constant-order policy is asymptotically optimal. However, the bounds proven there require the lead time to be very large before the constant-order policy becomes effective, in contrast to the good numerical performance demonstrated by Zipkin even for small lead time values. In this work, we prove that for the infinite-horizon variant of the same lost sales problem, the optimality gap of the same constant-order policy actually converges \emph{exponentially fast} to zero, with the optimality gap decaying to zero at least as fast as the exponential rate of convergence of the expected waiting time in a related single-server queue to its steady-state value. We also derive simple and explicit bounds for the optimality gap, and demonstrate good numerical performance across a wide range of parameter values for the special case of exponentially distributed demand. Our main proof technique combines convexity arguments with ideas from queueing theory.

math.PR↗

arXiv2016

Asymptotic optimality of Tailored Base-Surge policies in dual-sourcing inventory systems

Dual-sourcing inventory systems, in which one supplier is faster (i.e. express) and more costly, while the other is slower (i.e. regular) and cheaper, arise naturally in many real-world supply chains. These systems are notoriously difficult to optimize due to the complex structure of the optimal solution and the curse of dimensionality, having resisted solution for over 40 years. Recently, so-called Tailored Base-Surge (TBS) policies have been proposed as a heuristic for the dual-sourcing problem. Under such a policy, a constant order is placed at the regular source in each period, while the order placed at the express source follows a simple order-up-to rule. Numerical experiments by several authors have suggested that such policies perform well as the lead time difference between the two sources grows large, which is exactly the setting in which the curse of dimensionality leads to the problem becoming intractable. However, providing a theoretical foundation for this phenomenon has remained a major open problem. In this paper, we provide such a theoretical foundation by proving that a simple TBS policy is indeed asymptotically optimal as the lead time of the regular source grows large, with the lead time of the express source held fixed. Our main proof technique combines novel convexity and lower-bounding arguments, an explicit implementation of the vanishing discount factor approach to analyzing infinite-horizon Markov decision processes, and ideas from the theory of random walks and queues, significantly extending the methodology and applicability of a novel framework for analyzing inventory models with large lead times recently introduced by Goldberg and co-authors in the context of lost-sales models with positive lead times.

math.PR↗

arXiv2015

Second-order Markov random fields for independent sets on the infinite Cayley tree

Recently, there has been significant interest in understanding the properties of Markov random fields (M.r.f.) defined on the independent sets of sparse graphs. When these M.r.f. are restricted to pairwise interactions (i.e. hardcore model), much progress has been made. However, considerably less is known in the presence of higher-order interactions, which arise e.g. in the analysis of independent sets with special properties and the study of resource-constrained communication networks. In this paper, we further our understanding of such models by analyzing M.r.f. with second-order interactions on the independent sets of the infinite Cayley tree. We prove that the associated Gibbsian specification satisfies the celebrated FKG Inequality whenever the local potentials defining the Hamiltonian satisfy a log-convexity condition. Under this condition, we give necessary and sufficient conditions for the existence of a unique infinite-volume Gibbs measure in terms of an explicit system of equations, prove the existence of a phase transition, and give explicit bounds on the associated critical activity, which we prove to exhibit a certain robustness. For potentials which are small perturbations of those coinciding to the hardcore model at the critical activity, we characterize whether the resulting specification has a unique infinite-volume Gibbs measure in terms of whether these perturbations satisfy an explicit linear inequality. Our analysis reveals an interesting non-monotonicity with regards to biasing towards excluded nodes with no included neighbors.

math.PR↗

arXiv2014

Asymptotic Optimality of Constant-Order Policies for Lost Sales Inventory Models with Large Lead Times

Lost sales inventory models with large lead times, which arise in many practical settings, are notoriously difficult to optimize due to the curse of dimensionality. In this paper we show that when lead times are large, a very simple constant-order policy, first studied by Reiman (\cite{Reiman04}), performs nearly optimally. The main insight of our work is that when the lead time is very large, such a significant amount of randomness is injected into the system between when an order for more inventory is placed and when that order is received, that "being smart" algorithmically provides almost no benefit. Our main proof technique combines a novel coupling for suprema of random walks with arguments from queueing theory.

math.OC↗

arXiv2013

Steady-state $\mathit{GI}/\mathit{GI}/\mathit{n}$ queue in the Halfin-Whitt regime

We consider the FCFS $\mathit{GI}/\mathit{GI}/n$ queue in the so-called Halfin-Whitt heavy traffic regime. We prove that under minor technical conditions the associated sequence of steady-state queue length distributions, normalized by $n^{1/2}$, is tight. We derive an upper bound on the large deviation exponent of the limiting steady-state queue length matching that conjectured by Gamarnik and Momcilovic [Adv. in Appl. Probab. 40 (2008) 548-577]. We also prove a matching lower bound when the arrival process is Poisson. Our main proof technique is the derivation of new and simple bounds for the FCFS $\mathit{GI}/\mathit{GI}/n$ queue. Our bounds are of a structural nature, hold for all $n$ and all times $t\geq0$, and have intuitive closed-form representations as the suprema of certain natural processes which converge weakly to Gaussian processes. We further illustrate the utility of this methodology by deriving the first nontrivial bounds for the weak limit process studied in [Ann. Appl. Probab. 19 (2009) 2211-2269].

math.PR↗

arXiv2013

On the rate of convergence to stationarity of the M/M/N queue in the Halfin-Whitt regime

We prove several results about the rate of convergence to stationarity, that is, the spectral gap, for the M/M/n queue in the Halfin-Whitt regime. We identify the limiting rate of convergence to steady-state, and discover an asymptotic phase transition that occurs w.r.t. this rate. In particular, we demonstrate the existence of a constant $B^*\approx1.85772$ s.t. when a certain excess parameter $B\in(0,B^*]$, the error in the steady-state approximation converges exponentially fast to zero at rate $\frac{B^2}{4}$. For $B>B^*$, the error in the steady-state approximation converges exponentially fast to zero at a different rate, which is the solution to an explicit equation given in terms of special functions. This result may be interpreted as an asymptotic version of a phase transition proven to occur for any fixed n by van Doorn [Stochastic Monotonicity and Queueing Applications of Birth-death Processes (1981) Springer]. We also prove explicit bounds on the distance to stationarity for the M/M/n queue in the Halfin-Whitt regime, when $B<B^*$. Our bounds scale independently of $n$ in the Halfin-Whitt regime, and do not follow from the weak-convergence theory.

math.PR↗

arXiv2010

Non-Equilibrium Statistical Physics of Currents in Queuing Networks

We consider a stable open queuing network as a steady non-equilibrium system of interacting particles. The network is completely specified by its underlying graphical structure, type of interaction at each node, and the Markovian transition rates between nodes. For such systems, we ask the question ``What is the most likely way for large currents to accumulate over time in a network ?'', where time is large compared to the system correlation time scale. We identify two interesting regimes. In the first regime, in which the accumulation of currents over time exceeds the expected value by a small to moderate amount (moderate large deviation), we find that the large-deviation distribution of currents is universal (independent of the interaction details), and there is no long-time and averaged over time accumulation of particles (condensation) at any nodes. In the second regime, in which the accumulation of currents over time exceeds the expected value by a large amount (severe large deviation), we find that the large-deviation current distribution is sensitive to interaction details, and there is a long-time accumulation of particles (condensation) at some nodes. The transition between the two regimes can be described as a dynamical second order phase transition. We illustrate these ideas using the simple, yet non-trivial, example of a single node with feedback.

cond-mat.stat-mech↗