Constraint Management Uncertainty Estimation via Planning Separation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current constraint management systems face challenges in efficiently managing uncertainty and performing computational planning in data-dependent constraint networks, particularly in multidisciplinary analysis, due to intermixing of planning and computation, which limits design space exploration and increases computational complexity.
Innovation Solution
A method and system for estimating uncertainty in data-dependent constraint networks by propagating user selections through a bipartite graph, separating computational planning from numerical solution, and using well-formed formulas to determine applicable equations, allowing for efficient computational plans and uncertainty management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If computational planning and numerical solution are intermixed in constraint management systems, then the system can handle data-dependent constraints, but the computational complexity increases and design space exploration becomes limited
Solution Approach 1:
The patent segments computational planning from numerical solution into distinct phases. The computational planning phase determines the ordered sequence of computational steps using propositional forms and well-formed formulas, while the numerical solution phase executes these plans. This separation reduces computational complexity by avoiding repeated planning during numerical iterations, while maintaining adaptability through data-dependent propositional forms that guide the planning process.
2Productivity
If computational planning and numerical solution are separated, then computational efficiency improves and design space exploration expands, but the system becomes less flexible in handling data-dependent constraints
Solution Approach 1:
The patent introduces dynamics by making the computational plan adaptive through data-dependent propositional forms and well-formed formulas. The planning phase generates plans based on current data states, and these plans can be regenerated or modified as data changes during numerical solution. This dynamic approach maintains flexibility in handling data-dependent constraints while preserving computational efficiency through the separation of planning and solution phases.
3Reliability
If computational planning is performed repeatedly during numerical solution, then data-dependent constraints are accurately satisfied, but computational time increases
Solution Approach 1:
The patent applies preliminary action by performing computational planning before the numerical solution phase. The planning phase determines the complete ordered sequence of computational steps using propositional forms that capture data-dependent constraints. This preliminary plan is then executed during numerical solution without repeated planning, ensuring accuracy in satisfying constraints while minimizing computational time by avoiding redundant planning iterations.
Data Source
AI summary
Presented are rapid and robust techniques for estimating the uncertainty in product attributes (performance, cost, etc.) during the multi-disciplinary design and analysis phase of the product life-cycle. The techniques leverage the capabilities of a preexisting constraint management system that may be used to calculate performance and cost metrics of an engineering system as a function of the design structure and operational scenarios. The techniques are particularly useful when the constraint management system is used to automate the reverse computation required when the analyst specifies cost, schedule, or performance targets using approaches such as cost as independent variable. Disclosed techniques may also be applied to constraint management systems that include compound-valued variables.