Trade-offs, frontier, and sensitivity
The capacity frontier compares bounded capacity-multiplier cases and marks dominated modeled points: another modeled point uses no more aggregate capacity while producing at least as much modeled score. It is an exploratory view, not a complete mathematical Pareto frontier and not proof that unmodeled portfolios are dominated. Aggregate capacity combines cells for presentation; operational decisions still require unit-by-unit inspection.
Weight sensitivity varies the declared strategic-alignment, value, risk, and confidence policy and reports which initiatives enter or leave and how stable the selected set remains. It tests model dependence on those weights, not all uncertainty. Capacity estimates, demand, benefit, cost, interaction strength, eligibility, and category rules may need separate scenarios or specialist analysis.
Focus review on switch points and brittle selections. Ask whether a small defensible input change alters the portfolio, whether a protected initiative drives the result, and whether the model omits a consequential constraint. Sensitivity does not repair weak evidence; it shows where better evidence or a policy decision matters most.