OPTIMALLY CONFIGURING AN INFORMATION LANDSCAPE

• US 20140201116A1
• Filed: 03/25/2014
• Published: 07/17/2014
• Est. Priority Date: 05/09/2012
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Byrne, Brian P., Pandit, Sushain
• Granted Patent: US 9,245,224 B2
• Time in Patent Office: Days
• US Class Current: 706/19
• CPC Class Codes: G06N 3/08 Learning methods
• CPC Class Codes: G06N 3/084 Backpropagation, e.g. using gradient descent

Abstract

According to an embodiment of the present invention, a system optimizes an information processing environment, and comprises at least one processor. The system collects information pertaining to operational behavior of the information processing environment and including a plurality of parameters. A neural network structure is established to associate the parameters to a desired operational performance characteristic for the information processing environment. The neural network structure is trained with the collected information from the information processing environment to produce a model for the information processing environment. The model is optimized to determine values for the parameters and the information processing environment is adjusted based on the determined parameter values to attain the desired operational performance of the information processing environment. Embodiments of the present invention further include a method and computer program product for optimizing an information processing environment in substantially the same manner described above.

Claims

1. A method of optimizing an information processing environment comprising:

collecting information pertaining to operational behavior of the information processing environment, wherein the collected information includes a plurality of parameters for the information processing environment;

establishing a neural network structure to associate the parameters to a desired operational performance characteristic for the information processing environment;

training the neural network structure with the collected information from the information processing environment to produce a model for the information processing environment; and

optimizing the model to determine values for the parameters and adjusting the information processing environment based on the determined parameter values to attain the desired operational performance of the information processing environment.

2. The method of claim 1, further comprising:

collecting additional information pertaining to operational behavior of the adjusted information processing environment;

training the neural network structure with the additional information to update the model for the information processing environment; and

optimizing the updated model to determine updated values for the parameters and adjusting the information processing environment based on the determined updated parameter values.

3. The method of claim 1, wherein the neural network structure includes a plurality of neural networks each corresponding to one of the parameters, and an output node combining results of the plurality of neural networks to map the parameters to the desired operational performance characteristic of the information processing environment.

4. The method of claim 1, wherein the neural network structure includes plural sets of weights for application to information transferred between nodes of the neural network structure, and training the neural network structure includes:

determining values for the plural sets of weights that enable the neural network structure to associate parameter values with the desired operational performance characteristic, wherein the model includes the parameter values, determined sets of weights, and the desired operational performance characteristic.

5. The method of claim 4, wherein optimizing the model includes:

determining values for the parameters that optimize the desired operational performance characteristic based on the determined sets of weights.

6. The method of claim 1, wherein training the neural network structure includes:

producing training data for the neural network structure from the collected information from the information processing environment, wherein the training data includes values for the parameters and the desired operational performance characteristic.

7. The method of claim 1, wherein the information processing environment includes an information landscape employing a plurality of information processing systems each contributing to the desired operational performance characteristic of the information landscape.