Abstract: Aiming at the problem of power deviation caused by uncertain influence of renewable energy and

electric heating and cooling loads in the integrated energy system（IES）to realize integrated energy system optimal

dispatch，a double-layer IES optimal dispatch strategy was proposed by combining the stochastic model of chance

constrained programming（CCP）and the prediction model of back propagation neural network improved by

adaptive genetic algorithm（AGABP）. The upper layer used chance constrained programming to deal with the

uncertainty problem in the day-ahead optimal dispatch to relieve the problem of large power deviation caused by

the forecast error of renewable energy and multi-energy load，and the lower layer performed intra-day optimization based on the prediction model improved by adaptive genetic algorithm to correct the deviation of day-ahead optimal dispatch. To solve the problem that the stochastic optimization model is difficult to solve，the deterministic equivalence class method was used to transform the chance constraints in the model into deterministic constraints，and then the alternating direction multiplier method（ADMM）was used to optimize the improved model. Finally，an example simulation verified the effectiveness of the proposed scheduling strategy.

      Key words: integrated energy system（IES）；chance constrained programming（CCP）；back propagation

neural network；alternating direction multiplier method（ADMM）；double-layer optimal dispatch