constrained-inference

This paper proposes a constrained multi-source inference framework for accurate distribution system topology identification, using heterogeneous utility data and physical constraints. Tested on three feeders with over 8,000 AMI meters, it achieves over 95% reconstruction accuracy.