Hybrid Planning Engine for Tractable Probabilistic Contingencies
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional deterministic planning systems lack expressiveness in handling probabilistic actions and multiple possible worlds, while probabilistic planning systems face impractical complexities, making them untractable for current computer systems.
Innovation Solution
A hybrid probabilistic/deterministic planning system that uses a hybrid planning engine to convert deterministic partial plans into hybrid partial plans, splicing them into contingency plans to accommodate probabilistic actions and multiple possible worlds, and performs simultaneous reachability analysis to ground actions, incorporating a processor and memory for efficient computation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If conventional deterministic planning systems are used, then response speed is fast, but expressiveness is inadequate and probabilistic actions cannot be handled
Solution Approach 1:
The planning system is segmented into two distinct engines: a deterministic planning engine that handles fast computation and a probabilistic planning engine that handles expressiveness. Each engine operates independently on different aspects of the planning problem, allowing the system to maintain both speed and expressiveness without requiring a single engine to compromise on either.
2Adaptability or versatility
If conventional probabilistic planning systems are used, then expressiveness is improved and probabilistic actions are handled, but complexity becomes impractical and intractable
Solution Approach 1:
The system segments the complex probabilistic planning problem into manageable portions handled by the deterministic engine, while the probabilistic engine only handles the essential probabilistic aspects. This segmentation prevents the full complexity of probabilistic planning from overwhelming the system.
Solution Approach 2:
A hybrid contingency plan structure serves as an intermediary between the deterministic and probabilistic planning engines. This intermediate representation allows probabilistic actions to be expressed without requiring the full complexity of probabilistic planning, as the hybrid structure can accommodate probabilistic elements in a simplified manner.
3Device complexity
If conventional deterministic systems are used, then computational tractability is maintained, but probabilistic contingencies and multiple possible worlds cannot be accounted for
Solution Approach 1:
The system merges deterministic planning and probabilistic planning into a hybrid contingency plan that combines both approaches. The deterministic engine handles computationally tractable aspects while the probabilistic engine handles probabilistic contingencies and multiple possible worlds, achieving a combination of tractability and expressiveness.
Data Source
AI summary
A hybrid probabilistic/deterministic planning system including a hybrid planning engine responsive to a null hybrid contingency plan including user defined probabilistic problem including goals, possible effect actions, and the initial state of one or more possible worlds configured to extract a deterministic planning problem therefrom. A deterministic planning engine responsive to the hybrid planning engine configured to generate one or more deterministic partial plans. The d hybrid planning engine converts the one or more deterministic partial plans into one or more hybrid partial plans and splices the one or more hybrid partial plans into selected null hybrid contingency plans to generate one or more hybrid contingency plans for each agent in each of the one or more possible worlds.


