The Bethlehem Steel Sparrows Point plant produces 160" plates and 60" plates on their two mills. Their orders are classed into two major groups: NW, narrow widths, and BW, broad widths. The patterns used to fill customers' orders are cut from the original mother plates into patterns that are called mappings. The difficulty faced by the plant is how to manage the many variables that are encountered in the process and still accomplish the primary objectives:
- Minimizing the total trim loss for all the mother plates, and
- Cutting the fewest distinct order plate sizes from each mother plate.
The variables encountered are:
- Length of furnace time or run size,
- Milling time
- Size of original milled sheets ( length, width, and thickness),
- Order size and complexity of order,
- Scheduled hours of operation,
- Number and kinds of personnel needed to carry out the job, and
- Keeping orders together for ease in sorting and shipment.
A computer program was designed to manage these variables, and the many constraints placed upon them, in order to optimize the operation of the mill.
This program, or heuristic, makes a pattern for each day which takes the variable input for that day and schedules workers and orders so that the waste is minimized . At the end of the day, the following day's variables are input and the process is repeated. For example, if three customers each order a different plate size of the same thickness, the following two mapping scenarios of the customer orders into mother plates, might be considered:
- Each of the 30 mother plates contains plates from each of the three customers.
- The first 10 mother plates contain only order plates for the first customer, the second 10 mother plates contain only orders for the second customer, and so forth.
The heuristic for this example would attempt to:
- Cut as few surplus plates as possible (a productivity concern).
- Cut mostly high priority orders from the plate (a customer service concern).
- Cut orders from as few customers as possible from each plate (a logistical concern).
The new procedure is much faster than the original heuristic, particularly on more difficult problems. Its speed has made it possible to use it efficiently on a daily basis at the Sparrow Point plant. It consistently produces the expected results, and it also allows better use of the mother plate area which results in less waste and higher yield.