Study: Agents and science

Definition: Intelligent Agent (Wiki)

”In artificial intelligence, an intelligent agent (IA) is an autonomous entity which observes through sensors and acts upon an environment using actuators (an agent) and directs its activity towards achieving goals, being ”rational”. Intelligent agents may also learn or use knowledge to achieve their goals. They may be very simple or very complex. A reflex machine, such as a thermostat, is considered an example of an intelligent agent.”

Fig 1: Agents are useful for modeling economics, as an example (Turing Finance).

Agents are the most useful concepts used as abstractions and models in areas, where there are no concrete actors or dealers (Fig 1: Turing Finance). Agents can sell or buy or make any computations, using their knowledgebase.


Usecases: Agents are a widely accepted form of systems

System is abstraction for agent. So principles of systems (theory) and cybernetics can be applied for agents.

From the easiest to the most challenging possible tasks are:

  • Implementing, building, creating smthng
  • Analyzing
  • Planning
  • Learning
  • Negotiating

An agent often has an inference engine and knowledgebase as its features in the architecture.

It is typical for agents to communicate in networks and negotiate about deals, project conditions, to sell and buy. Agent is then an intelligent mediator, which is a popular design pattern in software.

Interdisciplinary research needs new definitions

Encapsulating science and practical disciplines is becoming important because of distributed working. Technological space means a representation system for models, defined in Model Driven Engineering, software science for modeling. Conceptual space is an equivalent for it. In research the equivalent is discipline, a context or problem space leading to solution space.

Fig 2: Technological space separates an homogeous area of software, which internally uses same kind of theory as a base.

This kind of thinking leads to meta science, understanding topologies of science. Wikipedia is an implementation for it, as well as Google Scholar.

If Fig 1 languages are defined by using grammars, GrammarWare and modeling happens in ModelWare, here by UML-technology. Hardware computes those programs.

SystemWare is then a conceptual space, which is dedicated in system based techniques (GoodReason). DocumentWare means methods to publish the information.

Example: Research as a conceptual spasce and an agent


Fig 3: Generic model for science and scientific tools.

Fig 3 shows GoodReason’s model for science  (the name rocket science” comes from the layout of the formulation). Resources are essential with their skills. The rows are Context (K), System and Physical (F).

The first column (1) is for seeking, the second is formulating (2) and the third is for apperception (findings).

  • The first line is for deepest science: K1, K2 and K3. K1 is for taxonomy and contect. K2 is for creativity to new formulation and K3 is for getting findings. This is not possible without use of relevant resources.
  • The second line is systemic. It needs systems thinking skills how to connect everything (S2). For it a conceptual framework, an architecture is needed to master the complexity (S1). As a result for that line a list of accepted and unaccepted criteria is gotten due to the viability tests.
  • The third line is for physical work, the mechanistic approach, what really happens in real life. F1 is the force to change something, which is done in a machine (F2), which produces an output (F3) to be used and evaluated in S2 and K2.
  • If the process is done completely, it is a possible candidate for improvements in practical life.

This simple model is rather generic and it can be used for product development and analyzing phenomena in a wide range, when definitions for symbols have been changed.

The same information can shortly be represented as a matrix with three rows and three colums.

OBS. The concept of Enterprise architectures (Zachman) contains the same three levels.


Links to the most relevant research

Systems theory as a science is introduced by New Mexico State University here.

Fig 3: Developing Systems Theory and Complexity Sciences.


In Fig 3 the subtheories and phases of development can be seen as cognitive models with their own intentionality and results based on their outputs on under conditions of each context.