Identifying internal parameters for planning is crucial to maximizing the performance of a planner. However, automatically tuning internal parameters which are conditioned on the problem instance is especially challenging. A recent line of work focuses on learning planning parameter generators, but lack a consistent problem definition and software framework. This work proposes the unified planner optimization problem (POP) formulation, along with the Open Planner Optimization Framework (OPOF), a highly extensible software framework to specify and to solve these problems in a reusable manner.

The solution of many continuous decision problem can be described as such a process: agent set out from the initial state, then go through a series of intermediate state and finally reach the goal state. Imagine an agent in a maze, which needs to find some key positions and pass through them one by one to get out. Agent has two types of behavior: one is the micro action taken at every state, which is similar to muscle activity, called reaction; another is the change of trend in reactions taken over a period of time, which is similar to thought of human, called planning [15]. For the agent in maze, reaction can be its every little moving step and planning can be its every determination of the position it should reach next. In a complicated scene with high-dimensional data stream, long-term decision process and sparse supervision signal, an agent trained only to react [9, 10] can hardly perform well (See Appendix A for demonstration). However, combining reaction and planning [3, 4, 14] can significantly improve its capability. The essence of such improvement is that agent has limited reaction capability and the introduction of planning releases agent from reacting in the whole task.

A plan with rich control structures like branches and loops can usually serve as a general solution that solves multiple planning instances in a domain. However, the correctness of such generalized plans is non-trivial to define and verify, especially when it comes to whether or not a plan works for all of the infinitely many instances of the problem. In this paper, we give a precise definition of a generalized plan representation called an FSA plan, with its semantics defined in the situation calculus. Based on this, we identify a class of infinite planning problems, which we call one-dimensional (1d), and prove a correctness result that 1d problems can be verified by finite means. We show that this theoretical result leads to an algorithm that does this verification practically, and a planner based on this verification algorithm efficiently generates provably correct plans for 1d problems.

A plan with rich control structures like branches and loops can usually serve as a general solution that solves multiple planning instances in a domain. However, the correctness of such generalized plans is non-trivial to define and verify, especially when it comes to whether or not a plan works for all of the infinitely many instances of the problem. In this paper, we give a precise definition of a generalized plan representation called an FSA plan, with its semantics defined in the situation calculus. Based on this, we identify a class of infinite planning problems, which we call one-dimensional (1d), and prove a correctness result that 1d problems can be verified by finite means. We show that this theoretical result leads to a practical algorithm that does this verification practically, and a planner based on this verification algorithm efficiently generates provably correct plans for 1d problems.