Form reasoning goals to ask itself to connect gaps in knowledge using other cases from memory
strategy integration is shifting from one strategy to another.
Ex. Metacognition chooses generate-and-test for all the strategies and determines the best, then moves to means-end analysis because it's best.
Metacognition and deliberation are overlapping
Metacognition helps w/ strategy selection
For each strategy it can select sub-tasks, recursively.
Represented by a two layer loop between deliberation and meta-cognition