Fig. 11: Simulation methodology flowchart for our recruitment algorithm.

First, prior knowledge about response distributions at each recruitment site is initialized (a). Then, after defining a target population and uniform distribution (b), the optimizer (c) determines the optimal recruitment resource allocation policy (d) among available sites. Recruitment is then simulated (e), yielding participants for the cohort (f) and updated knowledge about site distributions (a). This process is iterated until some goal is reached, such as participant count or number of steps.