Order Sequences and Order Release

For this system, sequences of orders for particular parts (batch size 1) are produced in the following way: for each machine, here is, for each shift, a prescribed maximal load level ζ is calculated as ζ (machine type, shift number) = ζ0 (machine type) + Δ ζ (shift number) Δ ζ is arbitrary parameter, the load variation.

Because of the processing sequences structure, the average load ζ0 can only be selected freely on three of the four categories of machines. The selected values are as:

For case 1: 98 percent on HV, DOER and 98 percent on ECOC

For case 2: 98 percent on HV, DOER and 70 percent on ECOC

The load amplitude Δ ζ is understood to be either 0, - 30 percent, or + 30 percent. The actual part mix for simulation is produced from a basic random part type sequence along with average element type frequencies of 10 percent for part 1 and 9 percent for other parts. For all shift, there is a load account for all machines. An order for an exact part type is accepted as extended as the load accounts for all needed machines and for this shift it must not exceed the maximum load level ζ for this shift and these machines. Also, this is skipped. If no more element types can be accepted for this shift, the order sequence for this shift is finish. The subsequent order in the basic sequence becomes the initial new order for the subsequent shift. The actual average part frequencies are hence various from the values in the basic sequence.

As load level can be above 100 percent machine capability, there are backlogs. They are carried above to the subsequent shift. Over the initial 5, 10, 15 and so on, consecutive shifts, the average load level variations are put to zero.

The real order sequences are produced from five various basic element type sequences and five fundamental load variation sequences for all of the conditions illustrated below.

In case 1, orders are released along with equivalent spacing over the duration of a shift within every 19 minutes. In case 2, all orders are released at the starting of a shift. This offers rise to a considerably longer queues for the initial half shift, in exacting at the loading station.

Hence, the characteristics of case 1 can be summarized as: evenly distributed high loads on the machines along with temporary overload; at the stations, short to moderate queues. Within case 2, the characteristics can be specified as: part types 1-5 only need machines that are not bottlenecks; only element types 6-11 utilize machines that may be overloaded; the queues are long; the loading station is a dangerous resource.

Additionally to above conversation, the due date distribution is a critical parameter also that affects the presentation of the  scheduling policies. The due dates are produced as multiples of four hour (1/2 shift) intervals from order free. The percentage of orders for every value of this demanded throughput time was found such as under scheduling as per to FIFO a prescribed value of percentage of late jobs of as 30, 50, 70, and 85% was attained. The resulting distributions are shown in Table no.2 (a) and (b).  The  resulting  ratios  of  demanded  to  real  throughput  times  for  the various situations under the FIFO rule have also been specified in aforementioned tables.

The conditions A-D correspond to case 1, the percentage of late jobs raising from 30-85, the conditions E-H correspond to case 2 along with similar levels of late jobs under the FIFO rule.

Table no.2 (a): Due Dates Distribution for Situation A-D