In the traditional mobile communication, voice and e-mail data occupied the large portion of traffic. However, in recent years, new mobile communication services, such as downloading movies and distributing various shop information with sound and animation, have appeared. These services require much more bandwidth to smoothly transmit multimedia data to users. Therefore, network capacity and transmission rate become more important. A cellular system is an infrastructure used in a mobile communication between a cellular phone, called a node, and a base station. In a cellular system, a base station can smoothly communicate with nodes in its communication area, called a cell, by avoiding overlap of frequency range with its adjacent cells. From the viewpoint of graph theory, however, that needs to divide the original frequency range into at least four sub-ranges. This leads to deteriorate the transmission rate. If a base station avoids overlapping its own cell with its adjacent cells by appropriately controlling its transmission power, that is perfect cell partitioning (PCP), it can use the whole of the original frequency range. In this thesis, we propose PCP for information distributing service of a shop. A shop may want to distribute advertising information to people in its own cell while preventing radio interference among adjacent shops. In this case, PCP is accomplished when a base station can adjust the size of the cell in which nodes do not exist in the vicinity of border among adjacent cells. Furthermore, we modify PCP to increase the success probability and the radio usage. In the modified PCP (MPCP), a base station sacrifices part of nodes around the border among adjacent cells. We first introduce PCP. Then, we provide geometrical analyses to clarify the feasible area of PCP. Through several simulation experiments, we verify the accuracy of the analysis and show the effectiveness of our proposal. Furthermore, we improve PCP for the purpose of increasing the effectiveness. Then, we show the effectiveness of modified PCP through simulation experiments. Simulation results show that PCP is effective when the number of nodes in a cell and the occupation ratio is lower than 7 and 1.5, respectively. On the other hand, MPCP is effective regardless of the number of nodes in a cell. Moreover, MPCP increases the success probability up to seven times higher than PCP when the occupation ratio is 1.2.

Akiko Miyagawa, Proposal and Evaluation of Perfect Cell Partitioning in Micro-Cellular Networks, Master's Thesis, March 2007.