Upper Bound Of System Capacity Research Articles

Traffic Load Optimization for Multi-Satellite Relay Systems in Space Information Network: A Proportional Fairness Approach.

Backbone satellites in a space information network (SIN) can be used as air base stations or data relay satellites (DRSs) to realize cross-system, cross-network and long-distance relay transmission. In this paper, a traffic load optimization problem for multi-satellite relay systems in SIN is considered to achieve highly efficient cooperative transmission and improve resource utility. A model of SIN based on a distributed satellite cluster (DSC) is considered, and the characteristics of the model are analyzed. Based on this, a hybrid resource management architecture combining distributed and central resources control schemes is proposed to realize a centrally controllable and distributed optimization of resources to meet various comprehensive service requirements. Two scenarios of multi-satellite relay systems in SIN are given, and traffic load optimization problems with joint bandwidth and power allocation for these two scenarios are formulated based on proportional fairness (PF) criterion to achieve traffic load balancing with considerable system capacity. The optimization problems in these two scenarios are proved to be a convex optimization problem with mathematical analysis, and the closed-form solutions of two problems in their dual domain are derived by dual transformation. With the closed-form solutions, two iterative algorithms based on the subgradient method are designed under the proposed hybrid resource management architecture to solve the problems in this paper. Simulation results show that the proposed schemes can effectively improve the upper bound of system capacity by resource sharing and cooperative relay, and it can balance the traffic load well with guarantees of a reasonable level system capacity compared with existing methods.

Using minimal cuts to study the system capacity for a stochastic-flow network in two-commodity case

Traditionally, the flow network is assumed to allow a single type of commodity to transmit through it. The system capacity is the maximum value of flow from the source to the sink. It is trivial that the system capacity is fixed for a deterministic flow network. However, for a stochastic-flow network (the capacity of each arc may have several possible values), the system capacity is not fixed. Hence, many authors proposed methods to calculate two performance indices, the probability that the system capacity is greater than d and the system capacity is less than d, for a level d in terms of minimal paths and minimal cuts, respectively. In the case that two types of commodities are transmitted through the stochastic-flow network, this paper uses the properties of minimal cuts to define the system capacity as a two-tuple vector and to propose an algorithm in order to evaluate a new performance index, the probability that the upper bound of system capacity is ( d 1, d 2), for a level ( d 1, d 2). Scope and purpose The max-flow min-cut theorem states that for a deterministic flow network, the maximum value of the flow from s to t equals the minimum capacity of all s– t minimal cuts. Hence, we can use the minimal cuts to deal with the flow problems in network analysis. This paper studies the stochastic-flow network in which two or more types of commodities are transmitted through it. From the point of view of quality management, it is an important issue to define the system capacity and then to develop some indices in order to evaluate the system performance. The existing performance index, the probability that the lower bound of the system capacity is ( d 1, d 2), for a level ( d 1, d 2) can be calculated in terms of minimal paths. The purpose of this paper is to use the properties of minimal cuts in order to define the system capacity for such a network. Then to develop a new performance index, the probability that the upper bound of the system capacity is ( d 1, d 2), for a level ( d 1, d 2) and to propose an algorithm to evaluate it. However, the two performance indices cannot convert into each other.