Research on Multi‐Objective Optimal Scheduling of Photovoltaic–Storage Flexible DC Microgrids Based on an Improved Multi‐Objective Northern Goshawk Algorithm

Photovoltaic–storage flexible DC microgrids require coordinated scheduling of economic operation and grid‐interaction stability. An improved multi‐objective northern goshawk optimisation method (IMONGO) is developed by extending the original northern goshawk optimiser to a Pareto‐based framework with an external archive and then introducing piecewise chaotic initialisation, a hybrid sine–cosine search strategy, and Cauchy mutation. The scheduling model minimises total operating cost and the standard deviation of tie‐line power subject to power‐balance, photovoltaic, micro gas turbine, battery‐energy‐storage, ramping, and grid‐exchange constraints. For ZDT1–ZDT4, the average inverted generational distance obtained by IMONGO is 2.365 × 10 − 3 , corresponding to reductions of 78.3%, 78.9%, and 86.0% relative to MONGO, MOGWO, and MOPSO, respectively; the average spacing indicator is reduced by 43.3%–49.2%. Relative to the direct MONGO baseline in the engineering‐derived microgrid case, the minimum operating cost decreases from CNY 8168.8 to CNY 8111.5. The cost‐conscious knee solution has an operating cost of CNY 8909.5 and a tie‐line‐power standard deviation of 397.1 kW. The distinction between Pareto‐front generation and compromise‐solution selection is clarified, and the effects of forecast uncertainty, tariff transferability, parameter provenance and grid‐exchange limits are discussed. The reported mean IGD and SP values indicate better average convergence and solution‐set uniformity on ZDT1–ZDT4. The engineering results demonstrate improvement relative to MONGO and a transparent trade‐off between operating economy and grid‐interaction smoothness; no broader claim of superiority over independent multi‐objective algorithms is made for the microgrid case.