Fr.: méthode du simplexe
An → algorithm for solving the classical → linear programming problem; developed by George B. Dantzig in 1947. The simplex method is an → iterative method, solving a system of → linear equations in each of its steps, and stopping when either the → optimum is reached, or the solution proves infeasible. The basic method remained pretty much the same over the years, though there were many refinements targeted at improving performance (e.g. using sparse matrix techniques), numerical accuracy and stability, as well as solving special classes of problems, such as mixed-integer programming (Free On-Line Dictionary of Computing, FOLDOC).