The theory of finite security markets developed by Harrison and Pliska [1] used the separating hyperplane theorem to establish the relationship between the lack of arbitrage opportunities and the existence of a certain martingale measure. In this paper we treat this theory by examining certain geometric properties of the sample paths of the price process, that is, we focus on the price increments of the stocks between one time period to the next and convert them to martingale differences through an equivalent change of measure. Thus, in contrast to Harrison and Pliska&s functional analytic derivation, our approach is based on probabilistic methods and allows a geometric interpretation which not only provides a connection to linear programming but also yields an algorithm for analyzing finite security markets. Moreover, we can make precise the connection between diverse expressions of economic equilibrium such as ‘absence of arbitrage’, ‘martingale property, and ‘complementary slackness property’.

We consider a Markov process with state space {0, 1}Z where Os become 1s irreversibly at rates which depend on whether none of a 0&s nearest neighbours (nucleation), its left-hand neighbour (right-hand growth), or its right-hand neighbour (left-hand growth) is in the 1-state. Furthermore, we assume that Os with both nearest neighbours in the 1-state remain in the 0-state forever and that at time 0 only the origin is in the 1-state. We determine the size distribution of the cluster (maximal sequence of 1s uninterrupted by Os) at the origin and the distribution of the time when its growth is stopped (by birth (nucleation) or competitive growth of neighbouring clusters). In the final state of the process, the spatial distribution of the (trapped) sites in the 0-state is considered. Some information on the size distribution of clusters well away from the cluster at the origin is obtained.

We consider the N-oscillator system of the van der Pol type, which contains a small positive parameter ε multiplying the non-linear damping and the random disturbance. For a formulation of the output we take the solution X(t) = (Xi(t))i=1···,N of the system of 2N-dimensional stochastic differential equations. Rotating each component Xi(t) about the origin of the plane by an angle t, we find that on time scales of order 1/ε together with sufficiently large N each Xi(t) behaves as the equi-ultimately bounded solution of an equation of the McKean type admitting a stationary probability distribution.

The method of Skorokhod embedding, that is to say of coupling to Brownian motion, is used to establish the asymptotic behaviour of one-dimensional scattering processes.

Let Xt, be a Markov random field assuming values in RM. Let In be a rectangular box in Zm with its center at 0 and corner points with coordinates ±n. Let (An) be a sequence of measurable subsets of RM such that neighborhood of t) → 0, for n → ∞; and let fn(x) be the indicator of An. Under appropriate conditions on the nearest-neighbor distributions of (Xt), the conditional distribution of given the values of Xs, for s on the boundary of In, converges to the Poisson distribution. An immediate application is an extreme value limit theorem for a real-valued Markov random field.

Properties of the conditional hazard rates of X1, · ··, Xn and Y1, · ··, Yn, which imply (X1, · ··, Xn) (Y1, · ··, Yn), are found. These are used to find conditions on the hazard rates of T = (T1, · ··, Tn) which ensure that T has the MIHR | property of Arjas (1981a) and the ‘weakened by failure’ property of Arjas and Norros (1984). Applications for load-sharing model and multivariate imperfect repair are given.

The family of distributions which have the lack-of-memory property is extended by incorporating simple patterns of ageing in the model. Some relatively simple multivariate distributions, obtained in this manner, might prove to be more realistic than distributions like the multivariate exponential.

Capacity expansion is the process of providing new facilities over time to meet rising demand. A general mathematical model of this process is presented, incorporating uncertain future demand (including the possibility of ‘surprises’), non-zero lead times and random cost overruns. In this model the decision maker controls the rate of investment in the current expansion project. Optimization is studied by methods of stochastic control theory. Numerical algorithms are presented which determine the optimal policy in some simple cases.

We study the problem of scheduling n given jobs on m uniform processors to minimize expected makespan (maximum finishing time). Job execution times are not known in advance, but are known to be exponentially distributed, with identical rate parameters depending solely on the executing processor. For m = 2 and 3, we show that there exist optimal scheduling rules of a certain threshold type, and we show how the required thresholds can be easily determined. We conjecture that similar threshold rules suffice for m > 3 but are unable to prove this. However, for m > 3 we do obtain a general bound on problem size that permits Bellman equations to be used to construct an optimal scheduling rule for any given set of m rate parameters, with the memory required to represent that scheduling rule being independent of the number of remaining jobs.

We prove a monotonicity result for the problem of optimal service rate control in certain queueing networks. Consider, as an illustrative example, a number of ·/M/1 queues which are arranged in a cycle with some number of customers moving around the cycle. A holding cost hi(xi) is charged for each unit of time that queue i contains xi customers, with hi being convex. As a function of the queue lengths the service rate at each queue i is to be chosen in the interval , where cost ci(μ) is charged for each unit of time that the service rate μis in effect at queue i. It is shown that the policy which minimizes the expected total discounted cost has a monotone structure: namely, that by moving one customer from queue i to the following queue, the optimal service rate in queue i is not increased and the optimal service rates elsewhere are not decreased. We prove a similar result for problems of optimal arrival rate and service rate control in general queueing networks. The results are extended to an average-cost measure, and an example is included to show that in general the assumption of convex holding costs may not be relaxed. A further example shows that the optimal policy may not be monotone unless the choice of possible service rates at each queue includes 0.

We consider the following circuit-switching problem which is one of the simplest possible extensions of the classical M/M/K/K model: there are p source centers connected to a hub by channels of various line-capacities and the hub is connected to a destination center by a common channel with its own line-capacity. A circuit requires a line from its source to the hub and another line from the hub to the destination. The holding times of circuits are independent, arbitrarily distributed random variables with means which depend on their source, and requests for circuits arrive in Poisson streams. Blocked calls are cleared. The problem is to calculate the blocking probabilities at each of the sources. The formal solution is well known but its calculation is exponentially hard in p.

We have developed a technique which extends recent results on integral representations and their asymptotic expansions to obtain the full expansion for the blocking probabilities in inverse powers of a large parameter N. The power of the method derives from two sources: first, an efficient recursive formula for calculating the general term of the expansion; second, tight upper and lower bounds which accompany the estimates. The computational complexity is polynomial in p. We report on computations.

This paper establishes a link between the Wiener–Hopf factorization and the phase-type method for studying the GI/G/1 queue. Using the Wiener–Hopf factorization, infinite-matrix type results are established for the GI/G/1 queue. An iterative numerical procedure (‘Levinson&s method’) based on these results is described. This method does not always converge. For the situation where either the interarrival times or the service times are of the so-called almost phase type (APH) an alternative, probabilistic derivation of the same results is given. This alternative derivation shows that in the APH situation Levinson&s method converges, converges essentially monotonically, and converges to the correct values.

The algorithm has been coded and examples of numerical results are included.

This paper studies the single-server queueing system with two independent input streams: a GI/G and an M/G stream. A new proof is given of an old result which shows how this system can be transformed into an equivalent ‘single input stream’ GI/G/1 queue, and methods to study that equivalent system numerically are given. As part of the numerical analysis, algorithms are given to compute the moments and the distribution function of busy periods in the M/G/1 queue, and of other related busy periods. Special attention is given to the single-server queue with independent D/G and M/G input streams.

This work is to be used in the modeling of real-time computer systems, which can often be described as a single-server queueing system with independent D/G and M/G input streams, see for example Ott (1984b).

A simple example is given of a point process with dependent inter-event intervals for which the distributions of the counting random variables Nt are given by the same formulas as for a renewal process. This example shows the care needed in establishing that a given point process has the renewal property, but also raises interesting questions about the class of point processes with given counting distributions.

Keilson and Servi introduced in [5] a variation of a GI/G/1 queue with vacation, in which at the end of a service time, either the server is not idle, and he starts serving the first customer in the queue with probability p, or goes on vacation with probability 1 – p, or he is idle, and he takes a vacation. At the end of a vacation, either customers are present, and the server starts serving the first customer, or he is idle, and he takes a vacation. The case p = 1, called the GI/G/1/V queue, was studied analytically by Gelenbe and Iasnogorodski [3] (see also [4]) and then by Doshi [1] and Fricker [2] who obtained stochastic decomposition results on the waiting-time of the nth customer extending the law decomposition result of [3]. Keilson and Servi [5] give a more complete analytic method of treating both the GI/G/1/V model and the Bernoulli vacation model: instead of the waiting time, they use a bivariate process at the service and vacation initiation epochs and the waiting-time distribution is computed as a conditional distribution of the above. In this note the law decomposition result is obtained from a reduction to the GI/G/1/V model with a modified service-time distribution just using the waiting time, with simple path arguments so that by [1] and [2] stochastic decomposition results are valid, which extend the result of [5].