Scheduling in egg packing plant

There are more than 350 million laying hens in the European Union. The laying hens produce close to 6.7 million tons of eggs each year. These eggs are collected from farms and centralized in packing stations where they are weighed and graded in weight classes Small, Medium, Large, Extra Large and packed into boxes destined for retailers. 

Grading is a fully continuous with the larger machines processing > 130.000 eggs per hour. As each egg is graded into a weight class, it needs to be individually sent to a packing lane where it is packed into a retail box. In packing plants, operators manually assigned orders to one or more packing lanes. This assignment needs to find a delicate balance in order to have sufficient receiving capacity in packing lanes for all the eggs in each weight class.

Eggs that cannot be sent to a packing lane because of insufficient capacity gets sent to the breaker where the yolk is separated from the egg white and sold separately. This fetches a much lower profit margin as is to be avoided as much as possible.

Operating packing lanes is costly since packing lanes, most of the time, require human operators. Therefore, the number of required packing lanes has a large impact on daily profitability of a packing station and should be minimized.

As in many packaging systems, certainly in food, packing lanes suffer often from packaging jams and breakdowns introducing downtime. Given the fact that it is so prevalent, any schedule should account for such breakdowns and ensure sufficient spare capacity for each weight class.

Since a hen lays nearly the same egg with regards to weight everyday, incoming supply from farms is relatively easy to forecast. Consequently, the egg supply follows a predictable weight distribution. Stochastic problems are typically hard to solve, nevertheless, this information on distribution should be exploited when designing a scheduling solution.