2. Comparing the list of dispatching rules and performance criteria from the surveyed   literature.

Buzacott and Yao [42] presented  a  comprehensive  review of the analytical models developed for the design and scheduling of FMSs. They strongly advocated analytical methods as giving a better insight into the system performance than simulation models. This point of view was adopted since, most probably, simulation techniques had not been  refined  up  that  time.  In the 1980s, there was less attention to the use of simulation in manufacturing applications [43], mainly because of the lack of model building expertise. Rahnejat [43] stated that analytical models are not efficient for reasonably sized problems. These models employ simplified assumptions that are  not  always valid in practice and also take a static view of the shop floor.

2.2 Statistics on Scheduling Problems and Performance Measures

2.2.1 Scheduling Problems

Table 1 summarises the scheduling problems that were con- sidered in the papers under  this  category.  Parts  dispatching was the most popular scheduling problem and machine selec- tion was the second one. These are long-lasting problems because  papers in  2000 were still considering  these  problems.

Table 1. Scheduling problem in general FMS scheduling    studies.

Scheduling Number of Reference  number Period

problem publications of publications

Parts dispatching 36 [4],  [23], [25], 1979–2000

[27],  [28], [32],

[34],  [36], [44],

[46–50], [50],

[52–54], [57–64],

[66], [70–77],

[79], [80]

Machine 8 [4,]  [45], [51], 1981–2000

selection [60], [67–69],

[81]

AGV  scheduling  4 [54], [67–69] 1986–1998

Others 5 [54–56], [66], 1986–1997

[76]

proposed rules have a significant influence on performance measures.

Stecke and Solberg [45]  carried  out  a  simulation  study of an FMS at the Caterpillar Tractor Company to show the impact of several machine sequencing rules on the performance of the FMS, under different loading objectives. The model contained ten machines with two carts to transport parts. They concluded that scheduling rules have a significant effect on the perform- ance of the FMS and some rules that were  known  to  be superior in a conventional job shop performed poorly in the FMS. They also demonstrated that the set of best-performing scheduling rules varied with the performance measures. Thus, there was no single scheduling rule that outperforms the others for  all  performance measures.

ElMaraghy  [25]  developed  a  general  discrete-event    FMS

  simulator.  The  package  was  programmed  in  FORTRAN and

was capable of simulating different configurations. Four dis- patching  rules  were  employed  including  SPT,  FOPR,  FIFO,

2.2.2