Approximations for structural response and design sensitivities significant
ly reduce the computational cost of structural optimization. To be useful,
the approximation must be accurate and computationally efficient. Damage to
lerant design puts even more stringent computational demands on optimizatio
n procedures and approximations due to the large structural changes inheren
t in damage. A robust approximation is evaluated and extended to include de
sign sensitivities. It is implemented in an optimization procedure and test
ed. It is shown that this approximation is accurate and efficient for damag
e tolerant optimum design of complex structural models. Last, a qualitative
measure is developed and tested that estimates the computational savings u
sing these approximations.