Surrogate modeling and control of medical digital twins

The paper formulates the same controlled Lotka–Volterra surrogate with removal controls −κ2Y and −κ3Z and targets “110% sheep, 50% wolves” relative to the original steady state, but resolves κ2, κ3 numerically (grid search/parametric surrogates) rather than proving analytic existence/uniqueness/feasibility conditions. It therefore does not state the necessary-and-sufficient inequalities ensuring a positive controlled equilibrium. The candidate solution correctly derives these conditions and the unique κ2, κ3 by solving the steady-state algebraic equations of the controlled model, given the nonnegativity of the removal rates and positivity of the state variables (from Eq. 2 and the stated control objective in the paper). Hence, the paper’s treatment is methodologically sound but incomplete with respect to this analytic point, while the model’s solution is correct. See the paper’s controlled ODE form and control setup (mechanistic LV + −κ2Y, −κ3Z) and the “50% wolves/110% sheep” target description and validation via grid search for ABM (black cross at κ2≈0.83%, κ3≈0.45%).