We consider a large-scale Ad-auction where adverts are assigned over a potentially infinite number of searches. We capture the intrinsic asymmetries in information between advertisers, the advert platform and the space of searches: advertisers know and can optimize the average performance of their advertisement campaign; the platform knows and can optimize on each search instance; and, neither party knows the distribution of the infinite number of searches that can occur. We look at maximizing the aggregate utility of the click-through rates of advertisers subject to the matching constraints of online ad allocation.

This paper describes a framework for admission control for a packet-based network where the decisions are taken by edge devices or end-systems, rather than resources within the network. The decisions are based on the results of probe packets that the end-systems send through the network, and require only that resources apply a mark to packets in a way that is load dependent. One application example is the Internet, where marking information is fed back via an ECN bit, and we show how this approach allows a rich QoS framework for flows or streams. Our approach allows networks to be explicitly analyzed, and consequently engineered.

This paper describes how some of the insights available from the stochastic analysis of dynamic routing may be incorporated into the classical mathematical programming approach to the design of networks. In particular, we present the results of a number of numerical investigations into network architectures for circuit-switched communication networks. Our investigations use recent theoretical results integrating network flow optimization and Markov decision processes to provide performance bounds for dynamic routing strategies. Following a tutorial introduction of the above mentioned topics we develop a sequence of network examples. Our first examples are familiar ones, such as symmetric fully connected networks and networks with moderate amounts of asymmetry, and we describe how network programming methods complement earlier work on dynamic routing. We then consider a variety of example networks which have a more sparse collection of links. These examples indicate the potential applicability of the methods to a variety of areas, including studies of the design, performance and resilience of future communication networks.>

This paper describes a simple and robust ATM call admission control, and develops the theoretical background for its analysis. Acceptance decisions are based on whether the current load is less than a precalculated threshold, and Bayesian decision theory provides the framework for the choice of thresholds. This methodology allows an explicit treatment of the trade-off between cell loss and call rejection, and of the consequences of estimation error. Further topics discussed include the robustness of the control to departures from model assumptions, its performance relative to a control possessing precise knowledge of all unknown parameters, the relationship between leaky bucket depths and buffer requirements, and the treatment of multiple call types.>

The authors investigate some of the consequences for dynamic routing schemes of dual- and multiparented networks, in which a call can enter (or leave) the network at two or more points. In particular, they compare bounds on the performance of optimal dynamic routing strategies which respectively ignore and utilize the multiparented structure, and show that simple schemes, easily implemented and analyzed, are able to achieve most of the additional advantages allowed to dynamic routing schemes by multiparenting. The robust behavior of these schemes under traffic mismatches as well as multiple link or node failure events is illustrated.>

Consider a closed cyclic queuing system consisting of M exponential queues.The Laplace-Stieltjes transform of the joint d~stribution of the consecutive sojourn times of a customer at the M queues is determined and shown to have a product form.The proof is based on a reverslbdity argument.