Co-Optimization of Resource Allocation and Operation in Large-Scale Energy Systems---Part I: Parallel-Distributed Theory

Integrated resource allocation across heterogeneous operators is imperative to achieve globally optimal planning in large-scale energy systems, yet faces fundamental barriers of computational intractability and stringent privacy protection requirements prohibiting sensitive data sharing. This paper develops a parallel-distributed framework resolving these dual challenges in large-scale planning-operation problems. Particularly, this complex problem is decomposed into two aspects: capacity optimization and location optimization. For the former, an enhanced equivalent projection (EP) technique handling mixed-integer optimization is employed, which compresses operator data into non-reconstructable proxies, reducing variables and ensuring privacy-essential security. For the latter, the location optimization leverages a tree-parallel Monte Carlo tree search (MCTS) algorithm that breaks commercial solvers’ resource utilization bottlenecks through asynchronous distributed computing, enabling previously intractable large-scale planning solutions. These innovations unify within a parallel-distributed framework where hierarchical planning center and operators coordination enables simultaneous privacy preservation through encrypted EP exchanges and computational acceleration via parallel MCTS processing. An illustrative example is presented at the end of this paper. In Part II, we deploy this methodology across integrated electricity-gas-heat systems and transmission-distribution networks, and verify its effectiveness in large-scale system planning.