On two-machine stochastic flow shop

Johnson solved the deterministic two-machine flow shop problem while Talwar found a solution for exponentially distributed job processing times. This paper shows that Johnson's and Talwar's solutions are satisfactory when the job processing time distributions (dfs) are "close to" degenerate dfs and exponential dfs, respectively. The closeness is examined in terms of the coefficient of variation of a df. It is also shown that the expected makespans derived from both solutions bound the expected makespan of all HNBUE distributed and independent job processing times.