In this paper we investigate the m-machine permutation flow shop scheduling problem where exact time lags are defined between consecutive operations of every job. This generic model can be used for the study and analysis of various real situations that may arise, for instance, in the food-producing, pharmaceutical and steel industries. The objective is to minimise the maximum lateness. We study polynomial special cases and provide a dominance relation. We derive lower and upper bounds that are integrated in a branch-and-bound procedure to solve the problem. Three branching schemes are proposed and compared. We perform a computational analysis to evaluate the efficiency of the developed method.