統合エネルギーシステムの低炭素経済運用：厳密実行可能な復号化フレームワークとRIMEアルゴリズムの次元補正

Unofficial AI-generated summary based on the public title and abstract. Not an official translation.

🔬研究者:Provides a rigorous optimization framework and a corrected RIME algorithm for low-carbon dispatch, useful for benchmarking and further algorithmic development.

🏢実務担当者:Offers a practical method for optimizing integrated energy systems under carbon constraints, potentially reducing operational costs while meeting emission targets.

🏛政策担当者:Highlights the limited abatement elasticity to carbon pricing in certain system configurations, informing the design of complementary policies.

The economic dispatch of integrated energy systems (IES) under carbon constraints couples electricity, heat and gas through numerous equality and inequality constraints, and the way those constraints are handled largely determines what a comparison between solvers actually measures. This work develops a strictly feasible decoding strategy that maps every vector in <inline-formula> <tex-math notation="LaTeX">$[{0,1}]^{D}$ </tex-math></inline-formula> to a dispatch schedule satisfying the power balances as exact identities: residuals are carried by explicit, priced variables representing grid export, renewable curtailment, heat rejection and unserved demand, and a reachability restriction enforces a cyclic state-of-charge condition by construction. Because the resulting model is a linear program under constant-efficiency assumptions, a solver-certified global optimum is available, and optimality gaps are reported rather than asserted. Using that reference we show that the coverage rate of the RIME algorithm is dimensionally misspecified: it governs how many coordinates are rewritten per candidate rather than the perturbation magnitude, so the published schedule <inline-formula> <tex-math notation="LaTeX">$E=\sqrt {t/T}$ </tex-math></inline-formula> drives the mutation order to nearly the full dimension late in the run, where the measured acceptance rate falls below 2%. Replacing it by the classical rate <inline-formula> <tex-math notation="LaTeX">$E=1/D$ </tex-math></inline-formula> is a single change with no tunable parameter. Over 30 independent runs at a common budget of 6000 objective evaluations, the resulting algorithm attains the lowest mean cost among twelve methods on two park-level systems, with optimality gaps of 3.51% and 3.29% and a significant margin over the runner-up (<inline-formula> <tex-math notation="LaTeX">$p=0.0002$ </tex-math></inline-formula> and <inline-formula> <tex-math notation="LaTeX">$p=0.023$ </tex-math></inline-formula>). An ablation conducted on a faithful baseline shows that the four-strategy combination provides no statistically significant benefit. Scenario analysis over a corrected ladder-type carbon trading formulation shows that total cost rises with the carbon price while emissions fall by only 0.86%, indicating limited abatement elasticity in a park whose thermal demand pins its cogeneration dispatch.

• semanticscholar https://doi.org/10.1109/access.2026.3720601first seen 2026-08-30 05:07:15 · last seen 2026-09-21 04:58:33