Often times it is necessary to simulate systems in order to define, refine, optimize or just get confidence in them before deployment. One approach is to connect a first principles or phenomenological modeling system to KSX and then interact with the model through KSX rules and neural models etc. Connecting KSX to third party software packages can be accomplished in a variety of ways and will be discussed in another blog post. Often times however these modeling systems don’t exist so another solution is needed.
KSX itself can also be used to simulate processes based upon rules and equations – from simple to complex. Agent based systems can be incorporated into a KSX based simulation which adds interesting qualities and potential to any KSX system.
Agent systems have been discussed a lot during the past few years. One definition provided by Wikipedia is “A multi-agent system (MAS) is a system composed of multiple interacting intelligent agents. Multi-agent systems can be used to solve problems, which are difficult or impossible for an individual agent or monolithic system to solve. Examples of problems, which are appropriate to multi-agent systems research, include online trading, disaster response, and modeling social structures.”
Again, using Wikipedia to further define qualities of agents we can see several of their important characteristics:
Autonomy: the agents are at least partially autonomous,
Local views: no agent has a full global view of the system, or the system is too complex for an agent to make practical use of such knowledge, and
Decentralization: there is no one controlling agent.
Typically, multi-agent systems research refers to software agents. And what is of real interest to us is the fact that agent systems can manifest self-organization and complex behaviors even when the individual strategies of all their agents are simple. In other words, systems of agents can manifest emergent and intelligent behavior.
In 2000 two associates, and I received a patent, US Patent 6112126 - Adaptive object-oriented optimization software system, which our original KnowledgeScape software was based. Reading the patent abstract it is easy to see that we were also talking about software agents or intelligent objects and all that makes them intelligent and how they can interact within a system.
KnowledgeScape, from the very beginning was designed in a way that it could be used to easily define an agent and then group them so that they can interact according to their simple rules. In my previous blog post I talked about intelligent traffic systems. The connection to multi-agent systems is the car, the traffic light and even pedestrians. Each can be defined as an agent with common rules and know-how.
In addition to the agents that might make up an Intelligent Traffic System, Swarm optimization, Ant optimization as well as Bee optimization are all examples of agent-based systems that can be implemented and experimented with in KSX.