Partially Preserving Action Orders in Top-Quality Planning
The authors propose a new computational problem called partially ordered top-quality planning, which allows specifying a subset of actions whose ordering in the plan is important, interpolating between the two extremes of considering all orders important or all orders unimportant.