Fuzzy call admission control combined with distributed dynamic channel assignment and reassignment for cellular mobile systems

Horng and Lin EURASIP Journal on Wireless Communications and Networking Fuzzy call admission control combined with distributed dynamic channel assignment and reassignment for cellular mobile systems Shih-Cheng Horng 0 1 Shieh-Shing Lin 1 2 0 Department of Computer Science and Information Engineering, Chaoyang University of Technology , 168 Jifong E. Rd., Wufeng District, Taichung City 41349 , Taiwan 1 Authors' information S-CH is currently an associate professor of the Department of Computer Science and Information Engineering at Chaoyang University of Technology , Taiwan , Republic of China. S-SL is now a professor of the Department of Electrical Engineering at St. John's University , Taiwan , Republic of China 2 Department of Electrical Engineering, St. John's University , 499, Sec. 4 Tam King Road, Tamsui, Taipei 25135 , Taiwan In a cellular mobile system (CMS), the service area is divided into cells, each of which has numerous channels, which are shared by two types of call - new calls and handoff calls. Giving a higher priority to handoff calls than new calls is common practice. However, giving too much priority to handoff calls will result to excessive blocking of new calls. This work firstly proposes a distributed dynamic channel assignment and reassignment (DDCAR) scheme to satisfy three types of constraint - co-channel constraint (CCC), adjacent channel constraint (ACC), and co-site constraint (CSC), simultaneously. The purpose is to minimize the number of available channels that become unavailable for the assignment to new calls and to maximize the number of unavailable channels that become available for release when calls complete. To provide a higher priority to handoff calls, a fuzzy call admission control (FCAC) scheme, combined with the DDCAR, is proposed herein for implementation at the base station. A 7 7 CMS with two traffic patterns is employed as the test example. The test results reveal that the FCAC scheme significantly reduces the dropping probability of handoff calls at the cost of increasing the blocking probability of new calls to an acceptable level. Call admission control; Cellular mobile system; Channel interference; Dynamic channel assignment; Fuzzy inference system; 7-cell reuse cluster system 1 Introduction Nowadays, the cellular mobile system (CMS) has become one of the most profitable mobile systems because more applications are being made available for smart phones, which also have very attractive functions [1,2]. One of the most important issues in the management of a CMS concerns the effective utilization of the limited resources of the radio spectrum and the maintenance of communication quality. To address this issue, channel assignment is required. In a CMS, three types of interference are generally treated as constraints; these are the co-channel constraint (CCC), the adjacent channel constraint (ACC), and the co-site constraint (CSC) [3,4]. The CCC states that any two cells within the channel reuse distance cannot use the same channel. The ACC implies that adjacent channels cannot be assigned to adjacent cells. The CSC is that any pair of channels in the same cell must be a specified distance apart. To maintain the quality of communication, channel assignment methods must incorporate the aforementioned constraints. The numerous channel assignment schemes include fixed channel assignment (FCA), dynamic channel assignment (DCA), borrowing channel assignment (BCA), and hybrid channel assignment (HCA) [5-9]. Among these, DCA schemes [8,9] are considered to be the most flexible and effective for improving the blocking probability. These schemes are divided into two categories - centralized schemes and distributed schemes [9]. In centralized DCA schemes, a channel is selected for a new call from a central pool of free channels, and a particular characterizing function is utilized to select one of the available free channels. In distributed DCA schemes, a channel is selected for a new call from the cell associated with the call or from interfering neighboring cells. Centralized DCA schemes theoretically provide near optimal performance. However, using a centralized approach to determine the best available channel for a new call without causing any channel interference is a combinatorial optimization problem, which is computationally intractable. Therefore, most existing centralized DCA schemes are heuristic. For instance, Lima et al. presented a genetic algorithm (GA)-based DCA method [10]; Kim et al. proposed a DCA method that is based on the GA and minimizes inter-cell interference [11]; Misra et al. developed a learning automata-based channel reservation scheme to find the optimal number of reserved channels in a system [12]; Zhao et al. proposed a DCA method that is based on a noisy chaotic neural network [13]; Krishna et al. presented a dynamic channel allocation scheme with efficient channel reservation for handoff calls [14]. In all of the above approaches, the DCA is centrally implemented in the mobile switching center (MSC) or in a cell with a central controller, which collects information about all of the cells in the CMS. The shortcomings of such centralized schemes are as follows; (i) the MSC is increasingly likely to fail owing to an overload of call requests as the size of the CMS grows with the growth in the number of mobile users [15,16], and (ii) determining the best available channel is time consuming even when heuristic methods are used, and a new call may be blocked as a result of just one unsuccessful assignment. These drawbacks suggest that distributed DCA schemes may be better for large CMS, because they assign and reassign channels at the base station of a cell [17,18]. Owing to the simplicity of the assignment algorithm at each base station, distributed DCA schemes are more attractive for implementation in a CMS. However, most current distributed DCA schemes impose only CCC. All current methods that deal with the CCC, ACC, and CSC simultaneously are centralized DCA schemes. To maximize channel utilization, most of centralized DCA schemes account for the packing condition and the resonance condition in selecting a channel to be assigned for a new call. The packing condition is applied to minimize the number of channels that are used whenever a call arrives; it allows the selection of those channels that are already in use in other cells as long as the three types of constraint are satisfied. The resonance condition assigns the same channels to cells within the reuse distance without causing any channel interference. In centralized DCA schemes, these two conditions are incorporated into the objective function over all cells that are governed by the MSC or are put into typical compact patterns for reference in the assignment process. In the real word, the place and time of the arrival of a new call is unpredictable. Accordingly, the resonance condition is too ideal to occur, and the assignment based on t (...truncated)