Finding an optimal sequence in terms of total completion time for processing ‘n’ jobs in ‘m’ processing centers in the shop floor is an important job for a shop floor supervisor. The problem is NP hard as the total number of sequences is (n!) for a permutation flow shop scheduling problem. Starting from Johnson’s Algorithm, many classical Heuristics were proposed in earlier days. After the invention of evolutionary algorithms and Meta Heuristics, finding an optimal sequence becomes comparatively easier. This paper deals with improving one such performance parameter, the make span, in a Permutation Flow Shop Scheduling problem in case of job idle times (though machines are available, jobs are not) observed in between two machines. The case of three machines, n jobs has been analyzed using the well known Taillard bench mark problems by coding in MATLAB R2008a and a new procedure has been proposed.