Abstract:In order to improve the computational efficiency,and to solve the problem that the dual gap cannot converge in the traditional Lagrangian relaxation algorithm in the optimization process,an improved Lagrangian relaxation algorithm is proposed to optimize the large-scale unit commitment problem.The method firstly optimizes the update path of Lagrange multipliers through the sine cosine algorithm to alleviate the oscillation,and secondly introduces the Cauchy mutation operator into the sine cosine algorithm to interfere with the current particle,so as to avoid falling into local optimum as much as possible and introduce adaptive weight update strategy to make particles approach the optimal solution faster.Finally,the power system of the units is used for simulation calculation,and the calculation results are compared with other algorithms.The results show that this method has advantages in calculation results and has practical application value.