Artificial Intelligence Constrained Multi-Objective Optimisation of Community Health Worker Household Visit Sequencing Under Variable Geographic Access Reporting: A computed Evaluation of Healthcare Detection Yield Equity

Community health worker (CHW) programmes in low- and middle-income countries allocate household visits under geographic access uncertainty, yet the equity consequences of alternative sequencing rules remain unexamined. the present analysis formulates CHW visit sequencing as a constrained multi-objective optimisation problem in which a planner maximises expected healthcare detection yield subject to an equity constraint on the distribution of visits across access strata, and in which the planner observes geographic access only through a noisy reporting channel whose reliability varies by stratum. A computed evaluation is conducted using the DHS Program indicator 'Discontinuation due to access/availability' (FP_RDIS_W_ACC), drawn from 74 country-survey observations across 28 African countries between 1992 and 2025 (mean 1.712, standard deviation 1.37, range 0.25 in Morocco to 4.90 in Lesotho), as the empirical anchor for access-stratum heterogeneity. The optimisation model is solved by a constrained Pareto frontier procedure that trades detection yield against a Gini-style equity penalty, and the reporting channel is modelled as a stratum-dependent misclassification process. The analysis derives conditions under which equity-constrained sequencing dominates unconstrained yield maximisation in detection-yield terms, and shows that the value of equity constraints increases monotonically with the dispersion of access-reporting reliability across strata. The contribution is a formal framework that makes the equity cost of access-reporting noise explicit and quantifiable, together with a set of propositions that specify when equity-constrained CHW sequencing is yield-improving rather than yield-sacrificing.