Derivative-free optimization(DFO)problems are optimization problems where the derivative information is unavailable.The least Frobenius norm updating quadratic interpolation model function is one of the essential unde...Derivative-free optimization(DFO)problems are optimization problems where the derivative information is unavailable.The least Frobenius norm updating quadratic interpolation model function is one of the essential under-determined model functions for model-based derivative-free trust-region methods.This article proposes derivative-free optimization with transformed objective functions(DFOTO)and gives a model-based trust-region method with the least Frobenius norm model.The model updating formula is based on Powell’s formula and can be easily implemented.The method shares the same framework with those for problems without transformations,and its query scheme is given.We propose the definitions related to optimality-preserving transformations to understand the interpolation model in our method when minimizing transformed objective functions.We prove the existence of model optimality-preserving transformations beyond translation transformations.The necessary and sufficient condition for such transformations is given.An interesting discovery is that,as a fundamental transformation,the affine transformation with a(non-trivial)positive multiplication coefficient is not model optimality-preserving.We also analyze the corresponding least Frobenius norm updating model and its interpolation error when the objective function is affinely transformed.The convergence property of a provable algorithmic framework containing the least Frobenius norm updating quadratic model for minimizing transformed objective functions is given.Numerical results show that our method can successfully solve most test problems with objective optimality-preserving transformations,even though some of such transformations will change the optimality of the model function.To our best knowledge,this is the first work providing the model-based derivative-free algorithm and analysis for transformed problems with the function evaluation oracle.This article also proposes the“moving-target”optimization problem as an open problem.展开更多
基金the National Natural Science Foundation of China(No.12288201)。
文摘Derivative-free optimization(DFO)problems are optimization problems where the derivative information is unavailable.The least Frobenius norm updating quadratic interpolation model function is one of the essential under-determined model functions for model-based derivative-free trust-region methods.This article proposes derivative-free optimization with transformed objective functions(DFOTO)and gives a model-based trust-region method with the least Frobenius norm model.The model updating formula is based on Powell’s formula and can be easily implemented.The method shares the same framework with those for problems without transformations,and its query scheme is given.We propose the definitions related to optimality-preserving transformations to understand the interpolation model in our method when minimizing transformed objective functions.We prove the existence of model optimality-preserving transformations beyond translation transformations.The necessary and sufficient condition for such transformations is given.An interesting discovery is that,as a fundamental transformation,the affine transformation with a(non-trivial)positive multiplication coefficient is not model optimality-preserving.We also analyze the corresponding least Frobenius norm updating model and its interpolation error when the objective function is affinely transformed.The convergence property of a provable algorithmic framework containing the least Frobenius norm updating quadratic model for minimizing transformed objective functions is given.Numerical results show that our method can successfully solve most test problems with objective optimality-preserving transformations,even though some of such transformations will change the optimality of the model function.To our best knowledge,this is the first work providing the model-based derivative-free algorithm and analysis for transformed problems with the function evaluation oracle.This article also proposes the“moving-target”optimization problem as an open problem.
