Rule-Based Plan Coordination Across Multiple Equipment Domains
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automated planning systems are inefficient and inflexible when dealing with multiple domains and sub-domains, as they lack coordination between different planning components, leading to inefficiencies and inability to react to changes in other plans.
Innovation Solution
A scalable infrastructure with a modular, hierarchically-structured architecture that enables communication through defined interfaces, allowing for multi-domain and multi-layered plan-based control, where a planner component generates plans and dispatches them to an execution engine, with continuous monitoring and re-planning capabilities to adapt to changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a monolithic approach is used to cover multiple drilling activities in a single domain model, then the system can maintain a unified plan, but the model becomes very unwieldy and difficult to maintain
Solution Approach 1:
The system divides the complex drilling domain into multiple sub-domains (drilling, mud management, casing, cementing) with separate domain models. Each sub-domain has its own planner that generates plans independently, then a coordination layer integrates these plans. This segmentation reduces the complexity of individual domain models while maintaining overall plan coherence through the coordination architecture.
2Device complexity
If a non-monolithic approach with separate domain models is used for drilling and mud management, then the system becomes more manageable, but the plans are inefficient and inflexible due to lack of coordination
Solution Approach 1:
The coordination layer continuously monitors the states and actions of all sub-domain plans, detecting conflicts and inefficiencies. When conflicts are detected (e.g., drilling plan actions that interfere with mud management plan actions), the coordination layer generates corrective adjustments and communicates them back to the affected sub-domain planners, who then modify their plans accordingly. This feedback loop ensures plan efficiency while maintaining manageability.
Solution Approach 2:
The coordination layer acts as an intermediary between separate domain models, translating and reconciling plans from different domains. It receives plans from drilling, mud management, and other sub-domains, identifies conflicts, and generates coordinated adjustments that ensure efficient execution across all domains without requiring a monolithic model.
3Ease of manufacture
If individual plans are generated for drilling and mud management separately, then each plan can be optimized for its specific domain, but the plans cannot react to changes in the other plan
Solution Approach 1:
The system implements dynamic plan coordination where the coordination layer continuously monitors plan execution and states. When changes occur in any sub-domain (e.g., drilling plan modifications), the coordination layer detects these changes and dynamically adjusts related plans in other domains (e.g., mud management plan) in real-time. This dynamic approach maintains ease of individual plan generation while achieving flexibility through continuous adaptation.
Data Source
AI summary
A recursive structure of inference rules is disclosed that enables translation of sensed data into high level predicates, e.g. to authorize the starts and ends of actions and detect violation of invariants, hence enabling coordination at different levels and/or across different domains at the same level within a multi-domain architecture. In embodiments, a processor may receive a condition configured to authorize an action of a system having equipment; identify a rule as a potential supporter of the condition in response to determining that the rule is associated with the condition; infer an implicant of the rule to be true when prerequisites of an antecedent of the rule are determined to be true; activate the rule in response to determining that the antecedent is true, wherein a validity of the condition is based on activation of the rule; and instruct the equipment to perform the action based on the validity of the condition.