91

Skip to main content
SHARE
Technology

Speculative Evaluation of What-if Scenarios Using Continuously Evolving Tree of Simulations on Finite Memory Machines

Invention Reference Number

202405650
Data stock image

Simulation cloning is a technique in which dynamically cloned simulations’ state spaces differ from their parent simulation due to intervening events. It is a powerful method to carry out what-if analysis by exploring and evaluating the impact of various permutations of intervening events. This technology is a novel, simulation cloning framework that executes a simulation cloning campaign capable of efficiently exploring an exponentially large space of clone simulations created by permutation of intervening events under a finite memory constraint.

Description

Due to the exponential growth in the number of possible clone scenarios, even for a small number of distinct intervening events, the practical efficacy of ensemble simulations performed using efficient simulation cloning method is severely limited by the maximum available memory of the computer. This technology, however, is a unique speculative simulation cloning framework that executes a simulation cloning campaign capable of efficiently exploring an exponentially large space of clone simulations created by permutation of intervening events under a finite memory constraint. It provides a theoretical analysis of the runtime characteristics of the approach, and has never-before-seen advantages, such as the memory management algorithm for as long as needed execution of exploratory simulations. 

Benefits

  • Greatly accelerates simulation computing
  • Drastically reduces computer run time for complex simulations
  • Extremely large “what if” scenarios become feasible
  • This capability does not exist; this is first example
  • Enables continuous execution of simulation cloning campaigns
  • Usability for artificial intelligence agent training, and many other real-world high-performance computing applications
  • Can be used for common computing, not just limited to HPC 
  • Users don’t need to know cloning specifics, details, or parallel execution details
  • Useful for digital twins, reinforcement learning

Applications and Industries

  • Almost any industry can benefit from this technology, including:
  • Power grid
  • Industrial systems
  • Cyber physical systems
  • Personalized cancer treatments
  • Manufacturing systems
  • Wildfire containment
  • Battlefield strategy evaluations
  • Vehicle traffic management
  • and more

Contact

To learn more about this technology, email partnerships@ornl.gov or call 865-574-1051.