I used Allen’s temporal interval calculus formalism for building a custom thing for my postdoc research back in the day. It’s more expressive than kanban though. More Gantt chart level.
There was a fair amount of work within/beyond the general STRIPS formalism (which came a decade after after GPS, out of SRI ~1971) into the 90s, and a model called HTN that goes beyond. But all hit complexity wall. Scheduling is even worse. But domain-specific heuristics do well.