A couple of relevant links:

• The action reward is a mini-framework that has been designed by Lokad to implement such prioritized ordering strategy. See https://docs.lokad.com/reference/abc/actionrwd.reward/
• The longer version of the topic through my lecture 6.1 at https://tv.lokad.com/journal/2022/5/12/retail-stock-allocation-with-probabilistic-forecasts/
• A discussion Why Safety Stock is Unsafe at https://tv.lokad.com/journal/2019/1/9/why-safety-stock-is-unsafe/ as a critic to the traditional perspective.