liveBook · Manning

Never switch policy

As explained in Applied Reinforcement Learning

A never switch policy continues producing the configured product until a target inventory is reached and then idles without changing setup.

A never switch policy avoids production-line setup costs by keeping the line configured for one product. It produces that product until its inventory reaches a target, then idles rather than switching to another product. section 2.5

Scooter and bike factory

If the line is set up for scooters, the policy makes scooters until their stock reaches the target, then idles. It does not switch to bikes, even when bike stock is low. section 2.5

Unlike an always-produce policy, this rule can idle when stock is sufficient. Unlike a setup-aware policy, it never switches to respond to low stock of the other product. A day-by-day simulation can compare policies over a planning horizon. section 2.5

How Applied Reinforcement Learning covers it

How deep it goes
Defines the target-inventory and idle decision rule and contrasts it with other heuristic policies.
Examples
Uses a scooter-and-bike factory and a day-by-day policy simulation.
What you can do after
Evaluate the policy against alternatives in a production-planning simulation.
How it connects
Connects production policies to inventory, setup costs, and Markov decision processes.

This explainer was generated by Manning's knowledge graph from the text of Applied Reinforcement Learning. The section links point to the parts of the book it draws on.