Driven by the "dual-carbon" strategic goals, the installed capacity of renewable energy sources, such as wind and solar power, has continued to expand. However, the grid integration and accommodation of high-penetration renewables intensify the peak-shaving and frequency-regulation pressure on conventional thermal power units, thereby increasing the operational costs and risks of the power system. Compared to traditional thermal units, multiple types of generation resources, including energy storage, pumped-storage hydropower, and virtual power plants (VPPs), offer superior flexibility and diverse control methods, providing robust support for peak and frequency regulation in modern power systems. Therefore, this paper proposes an optimal coordination strategy for day-ahead peak shaving and frequency regulation in power systems considering multiple generation resources. By optimizing the output of various resources, this method effectively reduces the peak-to-valley difference of the net load and alleviates the peak-shaving burden on thermal units, while ensuring adequate frequency-regulation reserves. Considering the distinct control characteristics of each resource, a bi-level optimization strategy is designed. The upper-level model, subject to the operational constraints of various generation resources and a permissible curtailment rate, aims to minimize the variance of the net load by coordinating flexible resources such as pumped-storage, battery storage, and VPPs to smooth the net load curve. The lower-level model then leverages the deep peak-shaving capability of thermal units. Subject to satisfying frequency-regulation reserve requirements, it minimizes the operational costs of thermal units by optimizing their generation schedules and reserve capacity allocations. Finally, simulation analyses based on a provincial power grid case study are conducted. The results demonstrate that the proposed strategy can effectively enhance the economic efficiency and stability of the power system's operation.