Tegel Foods is New Zealand's leading producer and supplier of poultry products. At the beginning of 2003 we offered to undertake a project for Tegelís Hornby processing plant in Christchurch. The Hornby plant operates five days out of every week, processing more than 50,000 chickens per day. Tegel was mainly concerned about the occurrence of plant idle time, this results when the stock of chickens to be processed runs out during the catchers shift.

The catchers are contracted by Tegel and are responsible for catching and transporting the chickens to the plant. At the time this project commenced, the contracted catchers felt that the available equipment was inadequate for the chicken catching and transporting operation. This includes not being able to complete the shift within a reasonable time, and having unnecessary plant idle time and catching delays at the farms. Although Tegel was willing to have the makeup of the fleet changed, they were unsure of which changes would make the most improvement to the catching and transporting operations performance.

       A scheduling tool named the VAS (Vehicle Assignment and Scheduling) tool was created to simulate and evaluate the effects of altering equipment levels. One of the challenges to effectively model the catching operation was to simultaneously schedule the trucks and catchers to the farms. This was necessary because the length of time that each truck stayed at the farms depended on the number of catchers catching at that farm. Although we found no literature on simultaneous scheduling that addressed a similar problem, an article by Hart et al. [3] approached the problem by splitting it into two stages. The two distinct parts that we split our problem into are below:

       Stage I,     Allocation of the van and trucks to the farms.

       Stage II,    Timing of the schedule, (when each vehicle arrives each destination).


       To model the assignment of trucks and vehicles to farms, the structure presented by Bailey [1] to assign nurses to hospital shifts was followed. Binary variables were used to represent the assignment of the trucks to the farms.

       The problem situation is complicated by parallel occurring events, in order to model the complex behaviour, the discrete event approach discussed by Luo et al. [4] was adapted. This specifically addressed:

a.         Generating multiple items at the same time

b.         Enforcing a specified processing sequence for items