Planning in real-world environments can be challenging for intelligent robots due to incomplete domain knowledge that results from unpredictable domain dynamism, and due to lack of global observability. Action language can be used for planning by formalizing the preconditions and (direct and indirect) effects of actions, and is especially suited for planning in robotic domains by incorporating defaults with the incomplete domain knowledge. However, planning with is very computationally expensive, especially when action costs are considered. We introduce algorithm PlanHG for formalizing domains at different abstraction levels in order to trade optimality for significant efficiency improvement when aiming to minimize overall plan cost. We observe orders of magnitude improvement in efficiency compared to a standard “flat” planning approach.