Nonlinear Programming Algorithms for Nonlinear Least Squares Adjustment by Parameters

Five feasible nonlinear programming algorithms dealing with nonlinear least squares adjustment by parameters are discussed. They are Newton method, speediest descending method, discrete Newton method, quasi-Newton method, and sequential quadratic programming method (SQPM). It is confirmed by analysis, comparison, and computation examples that SQPM is the most powerful tool to solve the problem of nonlinear least squares adjustment by parameters, without exactly computing the approximation of parameters.