Int. J. Production Economics 63 (2000) 217} 228 A study on the e!ect of the extent and location of protective capacity in #ow systems Sukran N. Kadipasaoglu, Wenuang Xiang, Simon F. Hurley, Basheer M. Khumawala* Center for Global Manufacturing, University of Houston, Houston, TX 77204-6282, USA Department of Technology, Massey University, Palmerston North, New Zealand Received 23 April 1998; accepted 23 November 1998 Abstract This research investigated in depth the e!ect of the level and the location of protective capacity relative to the location of the constraint through a simulation experiment. A production line with four stages is simulated. Location of the constraint, level of protective capacity, processing time variation and system down time are included as experimental factors. ANOVA is used for the analysis of the results, followed by Tukey's means test. The results show that protective capacity has a signi"cant e!ect on shop performance, however with diminishing returns. The location of the constraint resource also plays a signi"cant role in shop performance. The results also indicate that the extent of bene"ts derived from protective capacity or constraint location depends on the level of variability and system downtime.  2000 Elsevier Science B.V. All rights reserved. Keywords: Flow shop; Constraint resource; Protective capacity 1. Introduction Any manufacturing philosophy which claims to o!er speed and #exibility must address two phe- nomena: statistical #uctuations (system variability) and dependent events (dependency between the stages of production). Just-in-Time (JIT) and Drum-Bu!er}Rope (DBR) are two approaches which address these two phenomena, but in di!er- * Corresponding author. Tel.: #1-713-743-4721; fax: #1- 713-743-4940. E-mail address: bkhumawala@uh.edu (B.M. Khumawala)  Visiting scholar from Central South University of China, Changsa, China. ent ways. The JIT approach attempts to balance each stage of the production line. Since statistical #uctuations are an obstacle to successful JIT implementation, variability must be reduced to an absolute minimum. In addition, each stage in the production line may be allowed a limited amount of protective capacity to minimize the e!ects of the remaining variability. Krajewski et al. [1] found that protective capacity of up to 18% is allowed in some plants operating under the JIT philosophy. Besides protecting system output, pro- tective capacity reduces the need for bu!er inven- tory and, also, accommodates changes in the product mix as well as in customer orders. The time required to fully implement JIT systems, however, is signi"cant. 0925-5273/00/$ - see front matter  2000 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 5 - 5 2 7 3 ( 9 9 ) 0 0 0 2 0 - 1