Online convergence factor tuning for robust cooperative distributed economic dispatch

Solving economic dispatch problem (EDP) in a distributed way has attracted lots of attention in recent years due to its scalability and robustness to single points of failure. Robust distributed system Incremental Cost Estimation (RICE) algorithm has been proposed to solve the classic EDP in a distributed way considering communications information losses. However, assuring the stability of the algorithm without knowing the global information of the system is a challenging issue. This paper provides a distributed online approach to tune a certain parameter of the algorithm called “convergence factor” using only local information to assure the algorithm is stable. To do this, a local energy function is defined for each agent. As the algorithm proceeds, each agent uses a decaying mechanism to tune its convergence factor to ensure that its local energy function is within a certain bound. The summation of local energy functions represents an energy function for the entire network. Therefore, if each agent uses the tuning mechanism, the energy of the system would be forced to be constrained and the system will become stable. The effectiveness of the proposed approach is verified through several case studies.