Complex-Network Modeling of a Call Network |
| |
Abstract: | Recently, real-life data have revealed that the number of calls originating from or received by a telephone number in a network follows a power-law distribution. They show that a few telephone numbers make or receive a very large number of calls, whereas a large number of telephone numbers make or receive very few calls. The data have overthrown the general assumption that all telephone numbers are similar in generating telephone traffic. The first objective of this paper is to therefore construct a telephone call network (TCN) with connection properties following power-law distributions. With a more realistic TCN, researchers and engineers will be able to evaluate the telephone traffic behavior more accurately. Having constructed the aforementioned TCNs, we then consider the scenario when there is a sudden surge in the number of telephone calls, for example, during natural or man-made disasters. Under such a condition, the telephone network is usually overloaded and cannot operate properly. To mitigate the problem, we propose a preferential call blocking (PCB) scheme, aiming at blocking calls to target telephone numbers which have large numbers of incoming calls (in-strengths). We will investigate the effect on the carried traffic intensity when the PCB scheme is applied. We will compare the results with a benchmark, which corresponds to the case when all calls are blocked with equal probability. For the sake of completeness, we will also study the effectiveness of the blocking schemes when applied to a traditional TCN, in which all telephone numbers can call one another with equal probability. |
| |
Keywords: | |
|
|