Michael J. D. Powell

Born in London, Powell was educated at Frensham Heights School and Eastbourne College.[1] He earned his Bachelor of Arts degree followed by a Doctor of Science (DSc) degree in 1979 at the University of Cambridge.[7]

Powell was known for his extensive work in numerical analysis, especially nonlinear optimisation and approximation. He was a founding member of the Institute of Mathematics and its Applications and a founding Managing Editor of the Journal for Numerical Analysis. His mathematical contributions include quasi-Newton methods, particularly the Davidon-Fletcher-Powell formula and the Powell's Symmetric Broyden formula, augmented Lagrangian function (also called Powell-Rockafellar penalty function), sequential quadratic programming method (also called as Wilson-Han-Powell method), trust region algorithms (Powell's dog leg method), conjugate direction method (also called Powell's method), and radial basis function. He had been working on derivative-free optimization algorithms in recent years, the resultant algorithms including COBYLA, UOBYQA, NEWUOA, BOBYQA, and LINCOA. He was the author of numerous scientific papers[3] and of several books, most notably Approximation Theory and Methods.[8]