Comparing maintenance rules across several random seeds
A experimental plan sets out how maintenance rules in an abstract production line are compared across several random seeds.
Erlangen. The project Maintenance windows in small production chains investigates how different maintenance rules change the throughput of an abstract three-station line. Local maintenance decisions can lengthen queues elsewhere.
The experimental plan Experimental plan for maintenance rules governs the comparison. Each rule is run in a discrete-event simulation with documented failure assumptions across several random seeds. Throughput, waiting times, downtime and spread are assessed, not just a single favourable run.
In the scenario the project is ongoing. The plan describes what is to be examined and contains no results. It is a document; the production chain is abstract and does not replicate a real factory. Several random seeds are meant to prevent a rule from looking convincing merely because of one favourable course. The spread across runs is therefore reported alongside the averages.