What Are Intelligent Agents?

iRTDSS' DSIX decision support and information exchange intelligent agent software platform can run on a Windows Server, Workstation, or Windows IOT Enterprise operating system.

Within DSIX, each Intelligent Agent Process (IAP) is run as an independent Windows process, which is scheduled and launched by the system level DSIX Launcher process.

 Typically, each IAP will monitor a single source of data, interpret the data into information, and then send the resultant information to other systems, people, or IAPs.

 Within each IAP, over 90% of the code is typically supplied by pre-built code libraries, which deal with all the complexities of interfacing to different systems and handling resultant errors.

The actual task of interpreting the data retrieved from one or more systems is then performed by a Python script.

This script can be easily developed by business analysts, engineers, or IT people within the organization using DSIX because they are isolated from all the complexities of exchanging data with different systems.

Experience has shown that it is better to create many smaller IAP processes to perform limited tasks than one big IAP.

 This enables many IAPs to run in parallel on the new generation of Windows computer, where computers with 16 or more cores can mow be purchased for under $1,000.

For reliability, DSIX systems can be run on multiple computers, one at each geographically distributed site, so that operations can continue locally, even if the internet goes down.

In such a configuration IAPs at on site can exchange data with IAPs at another site, with messages reliably queued, with a store and forward mechanism, when the internet is not immediately available.

If needed DSIX can even be run on a mobile platform, as befits its military heritage.

While an IAP can use external Python libraries to perform complex tasks such as image recognition, experience has shown that they typically consist of less than a dozen rules in combination with relatively simple computations and algorithms, such as for exploring decision trees.

This makes it easy to create the Python scripts for IAPs, as they typically contain less than 100 lines of script code.

 Also, as DSIX supports dynamically changing Python scripts, while DSIX is running other IAPs, this enables a DSIX system to keep running 24x7 while changes are made to operational procedures.

This can be beneficial in these rapidly changing times .