A popular gantry-type placement machine includes several interconnected, autonomously
operating component placement modules and the machine was designed so as to able to
use different kinds of placement heads and vacuum nozzles in the modules, which can
be easily changed. Although this increases the flexibility of the production line,
the reconfiguration phases of the modules may be unproductive and one should keep
them to a minimum. In addition, the production times can be shortened by balancing
the workloads of the machine modules. Here, a two-step optimisation method for the
machine reconfiguration and workload balancing in the case of multiple Printed Circuit
Borad (PCB) batches of different sizes and PCB types is presented. The objective is
to minimise the total production time, and keep the machine configuration the same
for all batches. The proposed algorithm is iterative and it applies integer programming
for the workload balancing along with an evolutionary algorithm that searches for
the best machine configuration. In experiments, for single PCB types the proposed
algorithm obtained optimal or near optimal solutions. For multiple PCB types the solutions
favour the PCB types that have a bigger production time due to greater batch sizes,
but the total production time is still close to optimal.
