Karmarkar's algorithm: A threat to simplex? An elegant new method of solving linear equations has yet to be thoroughly tested

For almost 40 years, linear programming has been a stable discipline, marked by incremental improvements to the established standard algorithm, the “simplex method.” But in the summer of 1984, Narendra Karmarkar, a 30-year-old mathematician at AT&T Bell Laboratories, described an elegant and perhaps revolutionary new algorithm for solving linear programming problems, which he claimed was far faster than the simplex method.