Research

I study stochastic systems in which congestion, strategic behavior, uncertainty, and operational decisions interact. My work combines queueing theory and game theory with optimization, simulation, and data-driven methods.

Core methods: queueing theory · stochastic processes · game theory · optimization · simulation · machine learning

Queueing theoryGame theoryStochastic processesOptimizationSimulationMachine learning

01 / Behavior

Congestion, behavior & incentives

When users observe congestion and choose whether to join, wait, retry, switch, or leave, the system becomes a game—not only a queue.

What I study

I analyze how individual decisions respond to queue length, waiting costs, matching conditions, information, prices, and future utility. I am particularly interested in two-sided and matching systems where one population's behavior changes the incentives of the other.

Why it matters

Policies that look efficient under fixed demand may fail once users react. Equilibrium analysis helps identify when information, pricing, access rules, or capacity changes improve performance—and when they create new externalities.

02 / Performance

Stochastic service systems

How can we quantify delay, reliability, congestion, and failure when performance changes randomly over time?

What I study

My work uses continuous-time Markov chains, queueing models, conditional waiting-time analysis, distributional prediction, and stochastic simulation. The aim is not only to estimate an average, but to understand the mechanism that generates system-level performance.

Current direction

I am extending this line toward data-driven stochastic OR: combining structural models with real observations, testing policies under demand surges and prediction errors, and studying systems with exceptions, retries, and endogenous routing.

03 / Decisions

Decision support for complex operations

A mathematically attractive policy is useful only when it survives data limitations, operational constraints, uncertainty, and stakeholder interpretation.

What I build

I translate operational questions into computational frameworks using mathematical optimization, scenario analysis, forecasting, simulation, and reinforcement learning. Applications include inventory, infrastructure operations, traffic safety, manufacturing, and digital services.

Research principle

Prediction is an input to decision-making, not the endpoint. I focus on how forecasts enter policies, how errors propagate, which constraints are truly binding, and whether the result remains explainable enough to support action.

Research agenda. A central direction of my current work is data-driven stochastic operations research for service systems: integrating operating rules, user behavior, stochastic performance, and data so that policies can be evaluated under both normal and stress conditions.