Searcharxiv⌕ Search

arXiv subjects

Arpan Mukhopadhyay

Publications and source records attributed to Arpan Mukhopadhyay.

At least 19 recordsLinked to original sources
arXiv2026

Scaling Laws for Majority-based Opinion Dynamics in the Presence of Stubborn Agents

In a multi-agent system, there are often stubborn followers of specific opinions or beliefs. Motivated by this observation, in this paper, we aim to understand how stubborn agents affect the distribution of opinions in a network where both stubborn and non-stubborn agents interact with each other. To do so, we assume that all agents have an opinion in the set $\{0,1\}$ and each non-stubborn agent updates its opinion according to the $2k$\textit{-choices rule}, where the agent samples $2k$ neighbours (including both stubborn and non-stubborn neighbours) uniformly at random and adopts the majority opinion among the sampled group of neighbours and itself. We assume that a proportion of agents, $γ_i$, are stubborn followers of opinion $i\in \{0,1\}$. It is natural to expect that the steady-state distribution of the opinions in the network will be dominated by the opinion with the larger proportion of stubborn followers. We show that while this is true, the time to reach steady-state depends heavily on the values of the parameters $γ_0$ and $γ_1$. When the individual values of these parameters, as well as their difference, are small, it can take an exponentially long time (in the network size) to reach the steady-state. In sharp contrast, when at least one of the parameters $γ_0$ and $γ_1$ is large, the network reaches the steady-state in a time that is only logarithmic in the network size. Hence, there exists a sharp phase transition in the network dynamics based on the proportions of stubborn agents. We also characterise the behaviour of the system when the parameters $γ_0$ and $γ_1$ lie on the boundary of the phase transition. In this boundary region, we show using Stein's method that the dynamics are driven by a diffusion process which takes polynomial time to mix.

math.PR↗
arXiv2026

BOA Constrictor: Squeezing Performance out of GPUs in the Cloud via Budget-Optimal Allocation

The past decade has seen a dramatic increase in demand for GPUs to train Machine Learning (ML) models. Because it is prohibitively expensive for most organizations to build and maintain a large GPU cluster, organizations instead choose to rent GPUs from cloud providers. The customer is responsible for devising a policy for (i) deciding how many GPUs to rent at every moment in time to process a stream of ML training jobs and (ii) allocating the rented GPUs among the currently active jobs in the system. Because ML training jobs can be parallelized across different numbers of GPUs, the customer generally has many options for how many GPUs to use for each job. Allocating more GPUs to a single training job will cause the job to complete more quickly. However, the customer pays for each GPU-hour they use, and a training job receives a diminishing marginal benefit from running on additional GPUs. Hence, allocating too many GPUs to a single training job can dramatically increase the overall cost that the customer pays to the cloud provider. This gives rise to a cost-performance tradeoff that customers must balance when running training jobs in the cloud. To balance the cost-performance tradeoff, we develop BOA Constrictor, a new scheduler for ML training jobs which uses a Budget-Optimal Allocation (BOA) policy to squeeze the highest level of performance out of a cloud-deployed GPU cluster given a fixed budget constraint. We explicitly formulate the problem as a budget-constrained scheduling problem and derive the BOA policy which minimizes the average job completion time (JCT) of a stream of arriving jobs subject to the user's budget. For a given budget level, we demonstrate that BOA Constrictor can reduce average JCT by 1.6 times in small-scale implementation experiments and by 2 times in detailed, large-scale simulations compared to state-of-the-art heuristic based schedulers.

cs.DC↗
arXiv2026

Load Balancing with Individually Heterogeneous Servers

In this paper, we analyze the diffusion limit of Join-the-Shortest-Queue (JSQ) load balancing policies for a system with many parallel queues where each queue is being served by a server with a potentially different service rate. Prior asymptotic analyses of JSQ policies assumed servers to have service rates from a finite set which does not capture scenarios in modern data centers where service rates can vary at the level of individual servers. For systems with individually heterogeneous servers, tracking the empirical queue length distribution for each possible service rate becomes infeasible. To overcome this difficulty, we develop a framework based on measure-valued processes and provide a unified analysis of all JSQ-based load balancing policies under general tie-breaking rules. Our analysis identifies two key objects that distinguish the performance of different tie-breaking rules in the Halfin--Whitt regime, namely, the limiting fairness process which describes how idle servers are distributed across different service rates and the limiting routing measure which describes how arriving jobs are assigned across different service rates. In addition to characterizing the diffusion limits for different tie-breaking rules, we identify the tie-breaking rule which asymptotically minimizes the steady-state distributions of the diffusion-scaled total number of jobs and the diffusion-scaled number of waiting jobs. In proving these results, we develop crucial coupling-based sample-path comparisons which provide both policy-independent steady-state bounds and lower bounds to prove asymptotic optimality.

math.PR↗
arXiv2026

On Utility-optimal Entanglement Routing in Quantum Networks

Quantum networks are envisioned to enable reliable distribution and manipulation of quantum information across distances, forming the foundation of a future quantum internet. The fair and efficient allocation of communication resources in such networks has been addressed through the quantum network utility maximization (QNUM) framework, which optimizes network utility under the assumption of predetermined routes for competing user demands. In this work, we relax this assumption and aim to identify optimal routes that correspond to the maximum achievable network utility. Specifically, we formulate the single-path utility-based entanglement routing problem as a Mixed-Integer Convex Program (MICP). The formulation is exact when negativity is chosen as the entanglement measure for utility quantification or the network supports sufficiently high entanglement generation rates across demands. For other entanglement measures considered, the formulation approximates the problem with over 99.99% accuracy on evaluated real-world examples. To improve computational tractability, we propose a randomized rounding-based heuristic and an upper bound via the relaxation of the MICP. Furthermore, based on min-congestion routing, we introduce an alternative randomized heuristic and upper bound. This heuristic is computationally faster, while both the heuristic and the upper bound often outperform their counterparts on considered real-world networks. Our work provides the framework for extending classical flow-based and quality of service-aware routing concepts to quantum networks.

quant-ph↗
arXiv2026

Robustness of the 2-Choices Dynamics to Node Failures

In many applications, it becomes necessary for a set of distributed network nodes to agree on a common value or opinion as quickly as possible and with minimal communication overhead. The classical 2-choices rule is a well-known distributed algorithm designed to achieve this goal. Under this rule, each node in a network updates its opinion at random instants by sampling two neighbours uniformly at random and then adopting the common opinion held by these neighbours if they agree. For a sufficiently well-connected network of $n$ nodes and two initial opinions, this simple rule results in the network being absorbed in a consensus state in $O(\log n)$ time (with high probability) and the consensus is obtained on the opinion held by the majority of nodes initially. In this paper, we study the robustness of this algorithm to node failures. In particular, we assume that with a constant probability $α$, a node may fail to update according to the 2-choices rule and erroneously adopt any one of the two opinions uniformly at random. This is a strong form of failure under which the network can no longer be absorbed in a consensus state. However, we show that as long as the error probability $α$ is less than a threshold value, the network is able to retain the majority support of the initially prevailing opinion for an exponentially long time ($Ω(\poly(\exp(n)))$). In contrast, when the error probability is above a threshold value, we show that any opinion quickly ($O(\log n)$ time) loses its majority support and the network reaches a state where (nearly) an equal proportion of nodes support each opinion. We establish the above phase transition in the dynamics for both complete graphs and expander graphs with sufficiently large spectral gaps and sufficiently homogeneous degrees. Our analysis combines spectral graph theory with Markov chain mixing and hitting time analyses.

math.PR↗
arXiv2024

Asymptotic Optimality of the Speed-Aware Join-the-Shortest-Queue in the Halfin-Whitt Regime for Heterogeneous Systems

The Join-the-Shortest-Queue (JSQ) load balancing scheme is known to minimise the average response time of jobs in homogeneous systems with identical servers. However, for {\em heterogeneous} systems with servers having different processing speeds, finding an optimal load balancing scheme remains an open problem for finite system sizes. Recently, for systems with heterogeneous servers, a variant of the JSQ scheme, called the {\em Speed-Aware-Join-the-Shortest-Queue (SA-JSQ)} scheme, has been shown to achieve asymptotic optimality in the fluid-scaling regime where the number of servers $n$ tends to infinity but the normalised the arrival rate of jobs remains constant. {In this paper, we show that the SA-JSQ scheme is also asymptotically optimal for heterogeneous systems in the {\em Halfin-Whitt} traffic regime where the normalised arrival rate scales as $1-O(1/\sqrt{n})$.} Our analysis begins by establishing that an appropriately scaled and centered version of the Markov process describing system dynamics weakly converges to a two-dimensional reflected {\em Ornstein-Uhlenbeck (OU) process}. We then show using {\em Stein's method} that the stationary distribution of the underlying Markov process converges to that of the OU process as the system size increases by establishing the validity of interchange of limits. {Finally, through coupling with a suitably constructed system, we show that SA-JSQ asymptotically minimises the diffusion-scaled total number of jobs and the diffusion-scaled number of waiting jobs in the steady-state in the Halfin-Whitt regime among all policies which dispatch jobs based on queue lengths and server speeds.}

math.PR↗
arXiv2024

How to Rent GPUs on a Budget

The explosion in Machine Learning (ML) over the past ten years has led to a dramatic increase in demand for GPUs to train ML models. Because it is prohibitively expensive for most users to build and maintain a large GPU cluster, large cloud providers (Microsoft Azure, Amazon AWS, Google Cloud) have seen explosive growth in demand for renting cloud-based GPUs. In this cloud-computing paradigm, a user must specify their demand for GPUs at every moment in time, and will pay for every GPU-hour they use. ML training jobs are known to be parallelizable to different degrees. Given a stream of ML training jobs, a user typically wants to minimize the mean response time across all jobs. Here, the response time of a job denotes the time from when a job arrives until it is complete. Additionally, the user is constrained by some operating budget. Specifically, in this paper the user is constrained to use no more than $b$ GPUs per hour, over a long-run time average. The question is how to minimize mean response time while meeting the budget constraint. Because training jobs receive a diminishing marginal benefit from running on additional GPUs, allocating too many GPUs to a single training job can dramatically increase the overall cost paid by the user. Hence, an optimal rental policy must balance a tradeoff between training cost and mean response time. This paper derives the optimal rental policy for a stream of training jobs where the jobs have different levels of parallelizability (specified by a speedup function) and different job sizes (amounts of inherent work). We make almost no assumptions about the arrival process and about the job size distribution. Our optimal policy specifies how many GPUs to rent at every moment in time and how to allocate these GPUs.

cs.DC↗
arXiv2024

On Optimal Server Allocation for Moldable Jobs with Concave Speed-Up

A large proportion of jobs submitted to modern computing clusters and data centers are parallelizable and capable of running on a flexible number of computing cores or servers. Although allocating more servers to such a job results in a higher speed-up in the job's execution, it reduces the number of servers available to other jobs, which in the worst case, can result in an incoming job not finding any available server to run immediately upon arrival. Hence, a key question to address is: how to optimally allocate servers to jobs such that (i) the average execution time across jobs is minimized and (ii) almost all jobs find at least one server immediately upon arrival. To address this question, we consider a system with $n$ servers, where jobs are parallelizable up to $d^{(n)}$ servers and the speed-up function of jobs is concave and increasing. Jobs not finding any available servers upon entry are blocked and lost. We propose a simple server allocation scheme that achieves the minimum average execution time of accepted jobs while ensuring that the blocking probability of jobs vanishes as the system becomes large ($n \to \infty$). This result is established for various traffic conditions as well as for heterogeneous workloads. To prove our result, we employ Stein's method which also yields non-asymptotic bounds on the blocking probability and the mean execution time. Furthermore, our simulations show that the performance of the scheme is insensitive to the distribution of job execution times.

cs.DC↗
arXiv2023

On the Performance of Large Loss Systems with Adaptive Multiserver Jobs

In this paper, we study systems where each job or request can be split into a flexible number of sub-jobs up to a maximum limit. The number of sub-jobs a job is split into depends on the number of available servers found upon its arrival. All sub-jobs of a job are then processed in parallel at different servers leading to a linear speed-up of the job. We refer to such jobs as {\em adaptive multi-server jobs}. We study the problem of optimal assignment of such jobs when each server can process at most one sub-job at any given instant and there is no waiting room in the system. We assume that, upon arrival, a job can only access a randomly sampled subset of $k(n)$ servers from a total of $n$ servers, and the number of sub-jobs is determined based on the number of idle servers within the sampled subset. We analyze the steady-state performance of the system when system load varies according to $λ(n) =1 - βn^{-α}$ for $α\in [0,1)$, and $β\geq 0$. Our interest is to find how large the subset $k(n)$ should be in order to have zero blocking and maximum speed-up in the limit as $n \to \infty$. We first characterize the system's performance when the jobs have access to the full system, i.e., $k(n)=n$. In this setting, we show that the blocking probability approaches to zero at the rate $O(1/\sqrt{n})$ and the mean response time of accepted jobs approaches to its minimum achievable value at rate $O(1/n)$. We then consider the case where the jobs only have access to subset of servers, i.e., $k(n) < n$. We show that as long as $k(n)=ω(n^α)$, the same asymptotic performance can be achieved as in the case with full system access. In particular, for $k(n)=Θ(n^α\log n)$, we show that both the blocking probability and the mean response time approach to their desired limits at rate $O(n^{-(1-α)/2})$.

math.PR↗
arXiv2023

The Power of Two Choices with Load Comparison Errors

In this paper, we analyze the effects of erroneous load comparisons on the performance of the Po2 scheme. Specifically, we consider load-dependent and load-independent errors. In the load-dependent error model, an incoming job is sent to the server with the larger queue length among the two sampled servers with probability $ε$ if the difference in the queue lengths of the two sampled servers is less than or equal to a constant $g$; no error is made if the queue-length difference is higher than $g$. For this type of errors, we show that the benefits of the Po2 scheme is retained as long as the system size is sufficiently large and $λ$ is sufficiently close to $1$. Furthermore, we show that, unlike the standard Po2 scheme, the performance of the Po2 scheme under this type of errors can be worse than the random scheme if $ε> 1/2$ and $λ$ is sufficiently small. In the load-independent error model, the incoming job is sent to the sampled server with the {\em maximum load} with an error probability of $ε$ independent of the loads of the sampled servers. For this model, we show that the performance benefits of the Po2 scheme are retained only if $ε\leq 1/2$; for $ε> 1/2$ we show that the stability region of the system reduces and the system performs poorly in comparison to the {\em random scheme}.

math.PR↗
arXiv2023

Biased Consensus Dynamics on Regular Expander Graphs

Consensus protocols play an important role in the study of distributed algorithms. In this paper, we study the effect of bias on two popular consensus protocols, namely, the {\em voter rule} and the {\em 2-choices rule} with binary opinions. We assume that agents with opinion $1$ update their opinion with a probability $q_1$ strictly less than the probability $q_0$ with which update occurs for agents with opinion $0$. We call opinion $1$ as the superior opinion and our interest is to study the conditions under which the network reaches consensus on this opinion. We assume that the agents are located on the vertices of a regular expander graph with $n$ vertices. We show that for the voter rule, consensus is achieved on the superior opinion in $O(\log n)$ time with high probability even if system starts with only $Ω(\log n)$ agents having the superior opinion. This is in sharp contrast to the classical voter rule where consensus is achieved in $O(n)$ time and the probability of achieving consensus on any particular opinion is directly proportional to the initial number of agents with that opinion. For the 2-choices rule, we show that consensus is achieved on the superior opinion in $O(\log n)$ time with high probability when the initial proportion of agents with the superior opinion is above a certain threshold. We explicitly characterise this threshold as a function of the strength of the bias and the spectral properties of the graph. We show that for the biased version of the 2-choice rule this threshold can be significantly less than that for the unbiased version of the same rule. Our techniques involve using sharp probabilistic bounds on the drift to characterise the Markovian dynamics of the system.

math.PR↗
arXiv2023

Phase Transitions in Biased Opinion Dynamics with 2-choices Rule

We consider a model of binary opinion dynamics where one opinion is inherently 'superior' than the other and social agents exhibit a 'bias' towards the superior alternative. Specifically, it is assumed that an agent updates its choice to the superior alternative with probability $α>0$ irrespective of its current opinion and the opinions of the other agents. With probability $1-α$ it adopts the majority opinion among two randomly sampled neighbours and itself. We are interested in the time it takes for the network to converge to a consensus state where all the agents adopt the superior alternative. In a fully connected network of size $n$, we show that irrespective of the initial configuration of the network, the average time to reach consensus scales as $Θ(n \log n)$ when the bias parameter $α$ is sufficiently high, i.e., $α> α_c$ where $α_c$ is a threshold parameter that is uniquely characterised. When the bias is low, i.e., when $α\in (0,α_c]$, we show that the same rate of convergence can only be achieved if the initial proportion of agents with the superior opinion is above certain threshold $p_c(α)$. If this is not the case, then we show that the network takes $Ω(\exp(Θ(n)))$ time on average to reach consensus. Through numerical simulations we observe similar behaviour for other classes of graphs.

math.PR↗
arXiv2022

Asymptotic Optimality of Speed-Aware JSQ for Heterogeneous Systems

The Join-the-Shortest-Queue (JSQ) load-balancing scheme is known to minimise the average delay of jobs in homogeneous systems consisting of identical servers. However, it performs poorly in heterogeneous systems where servers have different processing rates. Finding a delay optimal scheme remains an open problem for heterogeneous systems. In this paper, we consider a speed-aware version of the JSQ scheme for heterogeneous systems and show that it achieves delay optimality in the fluid limit. One of the key issues in establishing this optimality result for heterogeneous systems is to show that the sequence of steady-state distributions indexed by the system size is tight in an appropriately defined space. The usual technique for showing tightness by coupling with a suitably defined dominant system does not work for heterogeneous systems. To prove tightness, we devise a new technique that uses the drift of exponential Lyapunov functions. Using the non-negativity of the drift, we show that the stationary queue length distribution has an exponentially decaying tail - a fact we use to prove tightness. Another technical difficulty arises due to the complexity of the underlying state-space and the separation of two time-scales in the fluid limit. Due to these factors, the fluid-limit turns out to be a function of the invariant distribution of a multi-dimensional Markov chain which is hard to characterise. By using some properties of this invariant distribution and using the monotonicity of the system, we show that the fluid limit is has a unique and globally attractive fixed point.

math.PR↗
arXiv2022

A Model of Job Parallelism for Latency Reduction in Large-Scale Systems

Processing computation-intensive jobs at multiple processing cores in parallel is essential in many real-world applications. In this paper, we consider an idealised model for job parallelism in which a job can be served simultaneously by $d$ distinct servers. The job is considered complete when the total amount of work done on it by the $d$ servers equals its size. We study the effect of parallelism on the average delay of jobs. Specifically, we analyze a system consisting of $n$ parallel processor sharing servers in which jobs arrive according to a Poisson process of rate $n λ$ ($λ<1$) and each job brings an exponentially distributed amount of work with unit mean. Upon arrival, a job selects $d$ servers uniformly at random and joins all the chosen servers simultaneously. We show by a mean-field analysis that, for fixed $d \geq 2$ and large $n$, the average occupancy of servers is $O(\log (1/(1-λ)))$ as $λ\to 1$ in comparison to $O(1/(1-λ))$ average occupancy for $d=1$. Thus, we obtain an exponential reduction in the response time of jobs through parallelism. We make significant progress towards rigorously justifying the mean-field analysis.

math.PR↗
arXiv2020

On the Throughput Optimization in Large-Scale Batch-Processing Systems

We analyze a data-processing system with $n$ clients producing jobs which are processed in \textit{batches} by $m$ parallel servers; the system throughput critically depends on the batch size and a corresponding sub-additive speedup function. In practice, throughput optimization relies on numerical searches for the optimal batch size, a process that can take up to multiple days in existing commercial systems. In this paper, we model the system in terms of a closed queueing network; a standard Markovian analysis yields the optimal throughput in $ω\left(n^4\right)$ time. Our main contribution is a mean-field model of the system for the regime where the system size is large. We show that the mean-field model has a unique, globally attractive stationary point which can be found in closed form and which characterizes the asymptotic throughput of the system as a function of the batch size. Using this expression we find the \textit{asymptotically} optimal throughput in $O(1)$ time. Numerical settings from a large commercial system reveal that this asymptotic optimum is accurate in practical finite regimes.

cs.PF↗
arXiv2020

Voter and Majority Dynamics with Biased and Stubborn Agents

We study binary opinion dynamics in a fully connected network of interacting agents. The agents are assumed to interact according to one of the following rules: (1) Voter rule: An updating agent simply copies the opinion of another randomly sampled agent; (2) Majority rule: An updating agent samples multiple agents and adopts the majority opinion in the selected group. We focus on the scenario where the agents are biased towards one of the opinions called the {\em preferred opinion}. Using suitably constructed branching processes, we show that under both rules the mean time to reach consensus is $Θ(\log N)$, where $N$ is the number of agents in the network. Furthermore, under the majority rule model, we show that consensus can be achieved on the preferred opinion with high probability even if it is initially the opinion of the minority. We also study the majority rule model when stubborn agents with fixed opinions are present. We find that the stationary distribution of opinions in the network in the large system limit using mean field techniques.

math.PR↗
arXiv2020

Storage Optimal Control under Net Metering Policies

Electricity prices and the end user net load vary with time. Electricity consumers equipped with energy storage devices can perform energy arbitrage, i.e., buy when energy is cheap or when there is a deficit of energy, and sell it when it is expensive or in excess, taking into account future variations in price and net load. Net metering policies indicate that many of the utilities apply a {customer selling} rate lower than or equal to the retail {customer buying rate} in order to compensate excess energy generated by end users. In this paper, we formulate the optimal control problem for an end user energy storage device in presence of net metering. We propose a computationally efficient algorithm, with worst case run time complexity of quadratic in terms of number of samples in lookahead horizon, that computes the optimal energy ramping rates in a time horizon. The proposed algorithm exploits the problem's piecewise linear structure and convexity properties for the \textit{discretization} of optimal Lagrange multipliers. The solution has a \textit{threshold-based structure} in which optimal control decisions are independent of past or future price as well as of net load values beyond a certain time horizon, defined as a \textit{sub-horizon}. Numerical results show the effectiveness of the proposed model and algorithm. Furthermore, we investigate the impact of forecasting errors on the proposed technique. We consider an Auto-Regressive Moving Average (ARMA) based forecasting of net load together with the Model Predictive Control (MPC). We numerically show that adaptive forecasting and MPC significantly mitigate the effects of forecast error on energy arbitrage gains.

eess.SY↗
arXiv2019

Optimal Storage Arbitrage under Net Metering using Linear Programming

We formulate the optimal energy arbitrage problem for a piecewise linear cost function for energy storage devices using linear programming (LP). The LP formulation is based on the equivalent minimization of the epigraph. This formulation considers ramping and capacity constraints, charging and discharging efficiency losses of the storage, inelastic consumer load and local renewable generation in presence of net-metering which facilitates selling of energy to the grid and incentivizes consumers to install renewable generation and energy storage. We consider the case where the consumer loads, electricity prices, and renewable generations at different instances are uncertain. These uncertain quantities are predicted using an Auto-Regressive Moving Average (ARMA) model and used in a model predictive control (MPC) framework to obtain the arbitrage decision at each instance. In numerical results we present the sensitivity analysis of storage performing arbitrage with varying ramping batteries and different ratio of selling and buying price of electricity.

eess.SY↗