On the variational process in parameter optimization (Q5903642)

It is shown that standard algebraic and analytic methods lead to the correct solution of the constrained optimization problem. The extra term in the second variation, discussed in a recent paper by \textit{D. G. Hull} [ibid. 56, 31-38 (1988; Zbl 0617.49002)], arises naturally from elementary calculus. It is present if and only if the variables of the objective function are dependent, but this is always the case for constrained optimization. Second and higher variations in the performance index, rather than in the augmented performance index, are considered.