초록 ▼

Techniques for tuning systems generate configurations that are used to test the systems to determine optimal configurations for the systems. The configurations for a system are generated to allow for effective testing of the system while remaining within budgetary and/or resource constraints. The co

Techniques for tuning systems generate configurations that are used to test the systems to determine optimal configurations for the systems. The configurations for a system are generated to allow for effective testing of the system while remaining within budgetary and/or resource constraints. The configurations may be selected to satisfy one or more conditions on their distributions to ensure that a satisfactory set of configurations are tested. Machine learning techniques may be used to create models of systems and those models can be used to determine optimal configurations.

대표청구항 ▼

1. A computer-implemented method of determining system configurations, comprising: receiving one or more initial configuration constraints related to a system;identifying a relationship between the one or more initial configuration constraints;determining at least one variable configuration constrai

1. A computer-implemented method of determining system configurations, comprising: receiving one or more initial configuration constraints related to a system;identifying a relationship between the one or more initial configuration constraints;determining at least one variable configuration constraint related to the system based at least in part on the identified relationship between the one or more initial configuration constraints;determining a set of configurations for the system based at least in part on the one or more initial configuration constraints and the at least one variable configuration constraint;generating a set of testing results by at least testing the validity of one or more configurations in the set of configurations;generating, based at least in part on the testing results, a model that defines at least one of the one or more initial configuration constraints and the at least one variable configuration constraint;determining, based at least in part on the generated model, an updated set of configurations that are valid based at least in part on the one or more initial configuration constraints and the at least one variable configuration constraint; andproviding the updated set of configurations for testing the system. 2. The computer-implemented method of claim 1, wherein generating the set of testing results includes operating the system under the one or more configurations. 3. The computer-implemented method of claim 1, further comprising: operating the system under the one or more configurations of the updated set of configurations;measuring performance of the operated system to generate a second set of testing results;creating, based at least in part on the second testing results, a second model by at least testing the system under the one or more configurations of the updated set of configurations; andusing the second model to determine at least one optimal configuration for the system. 4. The computer-implemented method of claim 1, wherein the updated set of configurations forms an optimal Latin hypercube. 5. The computer-implemented method of claim 1, wherein generating the model includes determining the at least one variable configuration constraint without prior knowledge of the variable configuration constraint. 6. The computer-implemented method of claim 1, wherein determining the updated set of configurations includes performing a particle swarm optimization algorithm to determine the updated set of configurations. 7. The computer-implemented method of claim 6, wherein the particle swarm optimization algorithm is a permutation particle swarm optimization algorithm. 8. The computer-implemented method of claim 1, wherein determining the at least one variable configuration constraint further comprises determining a set of values to be assigned to the at least one variable configuration constraint based on the identified relationship. 9. A computing system for generating configurations for a system under test, comprising: one or more processors; andmemory including executable instructions that, when executed by the one or more processors, cause the computing system to implement at least:a configuration validator configured to receive one or more initial configuration constraints related to a system and generate validity data that indicates whether one or more configurations in a set of configurations are valid based at least in part on the one or more initial configuration constraints;a trainer for training a model, configured to determine, based at least in part on the model and the generated validity data, at least one variable configuration constraint related to the system, the at least one variable configuration constraint determined based at least in part on identifying a relationship between the one or more initial configuration constraints; anda configuration generator configured to generate the one or more configurations for the system based at least in part on the one or more initial configuration constraints and the at least one variable configuration constraint to optimize one or more distribution metrics for the one or more configurations. 10. The computing system of claim 9, wherein the computing system is configured to operate the configuration validator, the trainer, and the configuration generator iteratively to determine the set of configurations for testing the system. 11. The computing system of claim 9, wherein the configuration validator includes a tester configured to test the system under the generated one or more configurations. 12. One or more non-transitory computer-readable storage media having collectively stored thereon instructions executable by one or more processors to cause a computer system to at least: receive one or more initial configuration constraints related to a system;identifying a relationship between the one or more initial configuration constraints;determine at least one variable configuration constraint related to the system based at least in part on the identified relationship between the one or more initial configuration constraints;determine a set of configurations for the system based at least in part on the one or more initial configuration constraints and the at least one variable configuration constraint;generate a set of testing results by at least testing the validity of one or more configurations in the set of configurations;generate, based at least in part on the testing results, a model that defines at least one of the one or more initial configuration constraints and the at least one variable configuration constraint;determine, based at least in part on the generated model, an updated set of configurations that are valid based at least in part on the one or more initial configuration constraints and the at least one variable configuration constraint; andprovide the updated set of configurations for testing the system. 13. The one or more non-transitory computer-readable storage media of claim 12, wherein generating the set of testing results includes operating the system under the one or more configurations. 14. The one or more non-transitory computer-readable storage media of claim 12, wherein the one or more computer-readable storage media further include instructions executable by the one or more processors to cause the computer system to further: operate the system under the one or more configurations of the updated set of configurations;measure performance of the operated system to generate a second set of testing results;create, based at least in part on the second testing results, a second model by at least testing the system under the one or more configurations of the updated set of configurations; anduse the second model to determine at least one optimal configuration for the system. 15. The one or more non-transitory computer-readable storage media of claim 12, wherein the updated set of configurations forms an optimal Latin hypercube. 16. The one or more non-transitory computer-readable storage media of claim 12, wherein generating the model includes determining the at least one variable configuration constraint without prior knowledge of the variable configuration constraint. 17. The one or more non-transitory computer-readable storage media of claim 12, wherein determining the updated set of configurations includes performing a particle swarm optimization algorithm to determine the updated set of configurations.