This dissertation studies three public transit operations control strategies with automatic vehicle location (AVL) data available. Specifically, holding control, stop-skipping control and vehicle dispatching with swapping are investigated. Moreover, AVL data from Tucson, Arizona are employed to investigate the methodologies for deriving vehicle operating parameters.The problem of holding vehicles at multiple holding stations can be modeled as a convex mathematical programming problem which can be solved to near optimality by a proposed heuristic. A simulation study on the holding problem suggests that holding control based on the proposed problem formulation can effectively reduce the total passenger cost. Also, multiple holding stations may offer more opportunities to regularize vehicle headways so that holding vehicles at multiple stations can further reduce the passenger cost compared to holding vehicles only at a single station.Stop-skipping is investigated to respond more rapidly to vehicle disruptions occurring in the middle of a route. Based on a preliminary analysis of the basic stop-skipping policy, a policy alternative is constructed. The stop-skipping strategy is formulated separately for both policies as a nonlinear integer programming problem. The problem solution relies on an exhaustive search method. Another simulation study is conducted to examine how the performance of the two policies change with the passenger distribution pattern, the vehicle disruption location and length, and the vehicle travel time variability. The simulation result suggests selective superiority of the two policies.The vehicle dispatching problem investigates the potential of integrating real-time swapping into the vehicle dispatching strategies at a transit transfer terminal. With a hypothetical study design, simulation is employed again to evaluate the significance of real-time swapping by comparing the performance of a swapping-holding combined strategy with the holding-only strategy. A sensitivity analysis is also employed to compare these two strategies among key transit operating factors.Finally, using three different understandings (assumptions) of vehicle operating behavior, regression methods are proposed for using AVL data to derive the vehicle running speeds and passenger boarding rates, which serve as inputs to the operations control models. The regression results show that the day-specific operating behavior may not be appropriate, and that operating behavior combining both trip-specific and day-specific effects seems to be slightly superior to the trip-specific behavior overall.