Dynamic resource allocation for epidemic control in multiple populations.

We develop a dynamic resource allocation model in which a limited budget for epidemic control is allocated over multiple time periods to interventions that affect multiple populations. For certain special cases with two time periods, multiple independent populations, and a linear relationship between investment in a prevention programme and the resulting change in risky behaviour, we demonstrate that the optimal solution involves investing in each period as much as possible in some of the populations and nothing in all the other populations. We present heuristic algorithms for solving the general problem, and present numerical results. Our computational analyses suggest that good allocations can be made based on some fairly simple heuristics. Our analyses also suggest that allowing for some reallocation of resources over the time horizon of the problem, rather than allocating resources just once at the beginning of the time horizon, can lead to significant increases in health benefits. Allowing for reallocation of funds may generate more health benefits than use of a sophisticated model for one-time allocation of resources.