With the advent of packet based networks and new applications in such networks, efficient use of network resources, congestion control and fair allocation of bandwidth become very important. Traffic scheduling algorithms provide a mean to reach these goals. Such algorithms determine the service order of packets of different sessions which are sharing an outgoing link. Two main objectives in designing scheduling algorithms are to reach fairness in bandwidth allocation and to reduce the worst case delay of packets. In addition, time complexity of the algorithms is also a very important factor, especially in broadband networks. The service time of a packet is very short in these networks and the algorithm must complete its run in such a short period. Hence, designing low complexity algorithms with desirable fairness properties and limited worst case delay of packets is the goal. This thesis tries to categorize different scheduling algorithms and study the properties of each category. Sorted-priority based scheduling algorithms as a main and important class of traffic scheduling algorithms is concentrated on and algorithms of this category are studied and compared in terms of time complexity, fairness and delay. Golestani's fairness index is used to compare their fairness properties and LR-servers theory is invoked to compare worst case delay of packets in these algorithms. Two main approaches are introduced to improve current algorithms; one is based on achieving lower fairness index and the other is based on reaching lower worst cast delay for packets. According to these approaches, two algorithms called LVT-SCFQ and FPFQ are introduced and their performances are evaluated in terms of fairness, delay and complexity.