Abstract: Despite its scalability and privacy advantages over centralized coordination schemes, decentralized price-based coordination in microgrids suffers from the demand response concentration effect, transferring flexible loads to low-price periods and yielding new demand peak, which hampers the efficient and secure operation of the system. Previous works have introduced auxiliary coordination parameters beyond electricity price to mitigate this effect. However, uniform values of these coordination parameters have been applied to all flexible loads, despite the effects of network constraints. To this end, this paper proposed an auxiliary parameters optimization for decentralized coordination in microgrids, applying optimal value for flexible loads in different nodes to mitigate demand response synchronization effect and minimize the total operational cost of the microgrid. Firstly, the decentralized coordination optimization model of microgrid and the demand response model of flexible loads, i.e. electric vehicles and smart appliances were established. Then a DRL-based approach to select the optimal values of auxiliary parameters was proposed, posing the parameter optimization problem in multi- dimensional continuous state and action spaces. Finally, simulation results demonstrated the effectiveness of the proposed optimization method.