||The phenomenal growth of data traffic, the increasing popularity of multimedia applications, and the convergences of mobile and Internet communications have fueled the development of the next generation (3G) wireless cellular networks capable of supporting multi-services. Various issues related to the next-generation wireless cellular system have to be carefully examined before such systems can be implemented. The fundamental challenge is to expand the single voice-centric service that the current wireless cellular network operates to provide a truly multi-services. This is further complicated by the limited radio resources are available. Therefore, traffic engineering is essential, which defines a set of policies and mechanisms that allow networks to effectively satisfy a diverse range of service requests. This includes bandwidth provisioning, call admission control, packet scheduling, and buffer management, flow control and etc. The main issue in traffic engineering is to strike a balance between the diversity and efficiency, as both are critical in providing the necessary service guarantee. This thesis focuses on the bandwidth allocation and call admission control, whose main objectives are somewhat conflicting with each other in that 1) the primary objective of bandwidth allocation schemes is to achieve maximal bandwidth efficiency by allowing more traffic in the system, while 2) the primary objective of call admission control is to limit the number of traffic that can be admitted into the system such that each individual traffic flow can obtain the desired service guarantee. We first study the problem of call admission control for two traffic types, namely voice and data. We answer two fundamental questions. First, how the bandwidth can be allocated to multiple traffic streams? Second, how to ensure that potentially different service requirement from diverse traffic can be guaranteed. We then investigate the bandwidth allocation for mixed traffic types. Specifically, we propose a new analytical model that can capture the performance in terms of call blocking and handoff dropping probabilities for two exemplary bandwidth allocation schemes, complete partition and complete sharing. We quantitatively demonstrate that both schemes can offer the needed service guarantee; the trade-off is that the complete sharing approach can obtain better bandwidth efficiency, while a partition based scheme can lead to easier derivation of the necessary control parameters. We next study the problem of bandwidth allocation for elastic data services, which is crucial for the next generation wireless cellular network operating over a packet-switching technology. We show that the proposed Dual Threshold Bandwidth Reservation (DTBR) scheme can effectively manage the bandwidth for voice with hard service guarantee, and for data with elastic service requirement. We proceed to study the problem of dynamic bandwidth allocation scheme, and extend to DTBR to a Dynamic Multi-Threshold Bandwidth Reservation scheme (DMTBR). We show that this scheme is capable to provide the service guarantee and differentiation for different type of traffic, while at the same being adaptive to changing traffic condition in the system. Finally, we also study the problem of bandwidth allocation for multi-services in a CDMA system. We compute the key performance for the downlink transmission, and demonstrate that the proposed bandwidth allocation scheme can also provide the necessary bandwidth and service provisioning for such systems.