Abstract: Optimized resource procurement and efficient scheduling strategies are pivotal in harnessing the full potential of demand response (DR) resources to provide ancillary services to the system. This paper studies the periodic purchase of DR resources and daily scheduling of peak-shaving, one of the most frequently occurring and influential ancillary service events in power system operation. First, the characteristics of peak-shaving events are clustered, and the probability is used to describe different decision behavior of the dispatching center. Then, the uncertainty model of periodic DR resource purchase is formulated and the optimal amount of purchased DR resources are obtained. Then, for intra-day scheduling, a multi-time scale energy and power coordination scheduling strategy based on rolling optimization is established with the forecasted load curve. Regulation energy is allocated on larger time scales to maximize the peak-shaving economy of the dispatching center, and regulation power is allocated on smaller time scales to optimize peak-shaving effects. Simulation results show that the periodic resource purchase mechanism can effectively improve the average peak-shaving economy of the dispatching center in a period. Meanwhile, the presented multi-time scale peak-shaving scheduling strategy help to make full use of purchased resources, and further achieve a win-win situation of both economic and effective peak shaving for the dispatching center.