[Objective] To reduce the total energy costs for commercial entities, satisfy renewable energy consumption mandates, improve the utilization rate of new energy, and alleviate reverse overload of distribution transformers, this study develops a configuration strategy for user-sited power sources including wind, solar and energy storage. The strategy accounts for time-of-use pricing curve, load guideline and load baseline demand response (DR). Furthermore, the constraints allowing reverse power transmission are incorporated, and the effectiveness of participating in three types of DR to minimize total energy costs is evaluated and compared. [Methods] A two-layer optimization model was established with the objective of minimizing total energy costs. This model involves alternating iterations between the configuration of user-sited power sources during the planning phase and the determination of the load baseline during the operational phase. The upper-level model considers the maximum values of multiple predicted new energy curves at any given time to establish a user-sited power source configuration model, and solves the capacity of new energy, energy storage, and nominal discharge time. The lower-level model calculates the mathematical expectation of total energy costs based on the outputs of the upper-level model and the probability of multiple predicted new energy curves. The mathematical expectation of transmission curves is used as the load baseline and fed back to the upper-level model. This iterative process continues until the results of two consecutive user-sited power source configuration and the load baseline are identical. By setting a reverse transmission constraint (not exceeding 20% of photovoltaic generation) and adjusting time-of-use pricing, benchmark incentive prices, and load guidelines, the influence on economic indicators was analyzed. A large commercial complex in a coastal province of China served as a case study for calculation and evaluation to verify the strategy's effectiveness from the perspective of reducing total energy costs. [Results] From the perspective of mathematical expectation, the proposed model reduced total energy costs by 2% to 10% compared to control schemes without user-sited power sources. The similarity when participating in baseline DR ranged from 98% to 100%, significantly higher than the 82% to 86% observed with guideline DR. Compared to guideline DR, participating in baseline DR reduced total energy costs by an additional 6% to 10%. Furthermore, the proportions of photovoltaic reverse transmission remained within 0% to 10%. [Conclusions] The strategy proposed in this paper satisfies the renewable energy consumption weights mandated provincial power grids while achieving high similarity between the load baseline and the transmission curve. The inclusion of reverse transmission constraints allows for an increased scale of user-sited power sources, significantly reducing the total energy costs for commercial users.