Artificial Intelligence Distributionally Robust Optimisation of Community Clinic Dialysis Chair Allocation Under Variable Water Supply Interruption Telemetry: A computed Healthcare Session Completion Equity Evaluation

Community dialysis units in water-scarce regions face a scheduling problem that conventional optimisation handles poorly: chair allocation must be decided before the day's water-supply interruptions are known, and the interruption telemetry that arrives during the day is both sparse and distributionally ambiguous. The model is formalised as a min-max problem over a Wasserstein ball of interruption distributions, reformulated as a tractable second-order cone programme, and evaluated in a computed healthcare setting that varies interruption frequency, telemetry reliability and equity weight. The evaluation is a design study, not an empirical trial: no patient records, clinic logs or telemetry streams were accessed. Across the computed scenarios the robust policy reduced the gap in session completion between the most and least served groups relative to a nominal expected-value policy, at a modest cost in mean completion, and the equity weight controlled the trade-off monotonically. The contribution is a formal bridge between distributionally robust optimisation and dialysis scheduling equity, together with a specification of the conditions under which telemetry-calibrated ambiguity sets outperform nominal allocation.