Artificial Intelligence Constrained Multi-Objective Optimisation of Community Clinic Temperature-Controlled Vaccine Storage Siting Under Variable Grid Outage Reporting: A computed Healthcare Spoilage Equity Evaluation

Temperature-controlled vaccine storage in community clinics depends on siting decisions that trade off spoilage risk against equity of access, yet the optimisation models used to support those decisions typically treat grid outage reporting as a fixed and reliable input. The framework couples a spoilage-risk model driven by ambient temperature exposure with an equity-weighted coverage objective and a budget constraint, and it is solved as a mixed-integer program over a computed clinic network. Ambient temperature inputs are taken from NASA POWER monthly reanalysis for Nairobi, Kenya, 2000-2023, comprising 312 monthly observations with a mean of 20.146 C, a standard deviation of 1.094 C and a warming trend of 0.2477 C per decade. The optimisation is run across a range of outage-reporting error regimes to characterise how misreporting shifts the Pareto frontier between spoilage reduction and equity. The contribution is a formal account of how reporting error enters the siting objective, together with a set of propositions identifying the conditions under which equity-optimal siting is robust to that error and the conditions under which it is not. The framework is proposed as a decision-support instrument; it is not an empirical evaluation of any deployed clinic network.