Leonardo Rojas-Nandayapa

Stability and performance of greedy server systems

Consider a queueing system in which arriving customers are placed on a circle and wait for service. A traveling server moves at constant speed on the circle, stopping at the location of the customers until service completion. The server is greedy: always moving in the direction of the nearest customer. Coffman and Gilbert conjectured that this system is stable if the traffic intensity is smaller than 1; however, a proof or counterexample remains unknown. In this review, we present a picture of the current state of this conjecture and suggest new related open problems.

Efficient tail estimation for sums of correlated lognormals

Our focus is on efficient estimation of tail probabilities of sums of correlated lognormals. This problem is motivated by the tail analysis of portfolios of assets driven by correlated Black-Scholes models. We propose three different procedures that can be rigorously shown to be asymptotically optimal as the tail probability of interest decreases to zero. The first algorithm is based on importance sampling and is as easy to implement as crude Monte Carlo. The second algorithm is based on an elegant conditional Monte Carlo strategy which involves polar coordinates and the third one is an importance sampling algorithm that can be shown to be strongly efficient.

Approximating the Laplace transform of the sum of dependent lognormals

Abstract Let (X 1,...,X n ) be multivariate normal, with mean vector 𝛍 and covariance matrix 𝚺, and let S n =e X 1 +⋯+e X n . The Laplace transform ℒ(θ)=𝔼e-θS n ∝∫exp{-h θ(𝒙)}d𝒙 is represented as ℒ̃(θ)I(θ), where ℒ̃(θ) is given in closed form and I(θ) is the error factor (≈1). We obtain ℒ̃(θ) by replacing h θ(𝒙) with a second-order Taylor expansion around its minimiser 𝒙*. An algorithm for calculating the asymptotic expansion of 𝒙* is presented, and it is shown that I(θ)→ 1 as θ→∞. A variety of numerical methods for evaluating I(θ) is discussed, including Monte Carlo with importance sampling and quasi-Monte Carlo. Numerical examples (including Laplace-transform inversion for the density of S n ) are also given.

Efficient simulation of finite horizon problems in queueing and insurance risk

Abstract Let ψ(u,t) be the probability that the workload in an initially empty M/G/1 queue exceeds u at time t<∞, or, equivalently, the ruin probability in the classical Crámer-Lundberg model. Assuming service times/claim sizes to be subexponential, various Monte Carlo estimators for ψ(u,t) are suggested. A key idea behind the estimators is conditional Monte Carlo. Variance estimates are derived in the regularly varying case, the efficiencies are compared numerically and also the estimators are shown to have bounded relative error in some main cases. In part, also extensions to general Lévy processes are treated.

Semiparametric Cross Entropy for Rare-Event Simulation

The Cross Entropy method is a well-known adaptive importance sampling method for rare-event probability estimation, which requires estimating an optimal importance sampling density within a parametric class. In this article we estimate an optimal importance sampling density within a wider semiparametric class of distributions. We show that this semiparametric version of the Cross Entropy method frequently yields efficient estimators. We illustrate the excellent practical performance of the method with numerical experiments and show that for the problems we consider it typically outperforms alternative schemes by orders of magnitude.

Non-existence of stabilizing policies for the critical push-pull network and generalizations

The push–pull queueing network is a simple example in which servers either serve jobs or generate new arrivals. It was previously conjectured that there is no policy that makes the network positive recurrent (stable) in the critical case. We settle this conjecture and devise a general sufficient condition for non-stabilizability of queueing networks which is based on a linear martingale and further applies to generalizations of the push–pull network.

Estimating tail probabilities of random sums of infinite mixtures of phase-type distributions

We consider the problem of estimating tail probabilities of random sums of infinite mixtures of phase-type (IMPH) distributions—a class of distributions corresponding to random variables which can be represented as a product of an arbitrary random variable with a classical phase-type distribution. Our motivation arises from applications in risk and queueing problems. Classical rare-event simulation algorithms cannot be implemented in this setting because these typically rely on the availability of the CDF or the MGF, but these are difficult to compute or not even available for the class of IMPH distributions. In this paper, we address these issues and propose alternative simulation methods for estimating tail probabilities of random sums of IMPH distributions; our algorithms combine importance sampling and conditional Monte Carlo methods. The empirical performance of each method suggested is explored via numerical experimentation.