《《2016 Recent developments in interior-point methods》.pdf

Chapter RECENT DEVELOPMENTS IN INTERIOR POINT METHODS Stephen J Wright Abstract The mo dern era of interiorp oint metho ds dates to when Kar markar prop osed his algorithm for linear programming In the years since then algorithms and software for linear programming have b ecome quite sophisticated while extensions to more general classes of problems such as convex quadratic programming semidenite programming and nonconvex and nonlinear problems have reached varying levels of ma turity Interiorp oint metho dology has b een used as part of the solution strategy in many other optimization contexts as well including ana lytic center metho ds and columngeneratio n algorithms for large linear programs We review some core developments in the area Keywords optimization interiorp oint metho ds INTRODUCTION Interiorpoint methods have been a topic of intense scrutiny by the optimization community during the past years Although methods of this type had been proposed in the s and investigated quite ex tensively during the s Fiacco and McCormick it was the announcement of an algorithm with intriguing complexity results and good practical performance by Karmarkar Karmarkar that ush ered in the modern era This work placed interiorpoint methods at the top of the agenda for a large and diverse body of researchers and led to a series of remarkable advances in various areas of convex optimization Today interiorpoint methods for linear programming have become quite mature both in theory and in practice and several highquality codes are available For the rival algorithm the simplex method the sudden appearance of credible competition spurred signicant improve ments in the software resulting in a quantum advance in the state of