To optimize renewable distributed generation in a distribution network, a bi-level optimization model is established considering the scheduling of controllable distributed generation and interruptible load. The upper level model aims to solve the location and capacity allocation of renewable distributed generation connected to the network with the goal of minimum voltage offset and minimum network loss in the distribution network system. The lower model solves the optimal output arrangement of controllable distributed generation and interruptible load in each period with the goal of minimum comprehensive operational cost of the distribution network. Scenario analysis is used to deal with the uncertainty in the optimization of renewable distributed generation. In order to reduce the amount of calculation, the K-means clustering algorithm is used to obtain typical scenes. Finally, an improved adaptive genetic algorithm is used to solve the model. The IEEE-33 node network is used as an example to analyze and calculate to verify the effectiveness of the proposed model and method. Results show that the method can effectively improve the applicability of the RDG optimization scheme in actual operation of the distribution network.