Complexity

Background

EITech offers customized development of multi-agent applications for administrations and businesses, primarily aimed at managing complexity.

EITech's founders developed interest in managing complexity in early 1980s when it became obvious to them that with the transition from Industrial to Information Society the social and economic environments within which we live and work will experience a rapid increase in complexity and that the new levels of uncertainty and dynamics of these environments will require new management and problem solving methods.

Fundamental Features of Complexity

The following paragraphs from Wikipedia are a good introduction to the concept of complexity.Complexity has always been a part of our environment, and therefore many scientific fields have dealt with complex systems and phenomena. Indeed, some would say that only what is somehow complex what displays variation without being random is
worthy of interest.

The use of the term complex is often confused with the term complicated. To understand the differences, it is best to examine the roots of the two words. "Complicated" uses the Latin ending "plic" that means, "to fold" while "complex" uses the "plex" that means, "to weave." Thus, a complicated structure is one that is folded with hidden facets and stuffed into a smaller space. On the other hand, a complex structure uses interwoven components that introduce mutual dependencies and produce more than a sum of the parts... This means that complex is the opposite of independent, while complicated is the opposite of simple.

A Definition of complexity

For the purposes of solving complex problems described here, complexity is defined as follows.

A situation, problem or system is said to be complex if it exhibits the following characteristics:

1. It consists of a large variety of autonomous components, which may have conflicting goals, and are engaged in rich interaction; there is no centralized control; a good example of a complex system is the global market, in which massive numbers of different suppliers, customers, middlemen, investors and administrators, each with individual declared or undeclared objectives, produce, sell, purchase, borrow, invest and regulate, often exchanging information using high-speed communication networks such as the Internet.
   

2. Its behavior emerges from a myriad of interconnected local behaviors of constituent components; it is therefore unpredictable; for example, the matching of supplies to demands in the global market emerges from billions of transactions between individual agents.
   

3. It behaves "far from equilibrium", the expression used by Prigogine, or "at the edge of chaos", as described by Santa Fee Institute researchers, because it is frequently disturbed and has no time to settle down in intervals between two events that cause disruptions. For example, at present the occurrence of events that affect distribution of supplies to demands is so frequent that, contrary to conventional economic theory, free market has never sufficient time to reach equilibrium.
   

4. Its behavior is nonlinear; a small disturbance may cause large changes in their global behavior (butterfly effect) whilst large disturbances may go unnoticed. A butterfly effect (or snowball-like behavior) causes unpredictable serious consequences in many complex systems such as global financial networks (consider current sub-prime credit fiasco), self-combustion and epidemics (e.g., aids). Large unpredictable changes in global behavior of complex systems are sometimes called Black Swans.
   

5. It is capable of autonomously self-organizing, that is, changing its behavior or even structure in response to unpredictable events, and is therefore adaptive to changes and resilient to attacks.
   

A Method for Managing Complexity

Contrary to expectations of many protagonists of old-style physical sciences, complexity cannot be simplified and the behavior of complex systems cannot be predicted. Mathematical models that leave out diversity of constituent components of a complex situation or ignore conflicting goals of constituent actors lead to misleading results. Deriving clever formulae to fit experimentally observed trajectories have no practical value because history of a complex system is no a guide for its future behavior. Conventional mathematical optimizers and schedulers that are effective when markets are reasonably stable, and which required 8 to 10 hours to plan a big business process such as a car production plant, become useless when the complexity of markets grows and the frequency of changes in demand increases to once in every two hours.

Services and manufacturers that sell to the global market are facing levels of complexity never experienced before, and, as Thomas Kuhn predicted, new problems are driving the emergence of a new paradigm.

The key assumption of the new paradigm is that although the behavior of complex systems is emergent and therefore cannot be predicted, it can be managed. The well-managed complex systems exhibit controlled emergence, in other words, their behavior although unpredictable, is always within prescribed boundaries.

Tools for managing complexity

Tools for the management of complexity contain two main components:

1. 1.Ontology, which contains the problem domain knowledge represented as a network, where nodes are classes of objects, characterized by attributes and rules of behavior, and links are relations between object classes.
   

2. 2.A Virtual World, which is a model of the domain of the real world that is being managed. Using classes of objects from ontology we build a Scene, which is a network, in which nodes are virtual instances of objects and links are their relations. As real events occur the network of instances, the scene, changes. A set of autonomous Agents, each assigned to an object instance, exchange messages; the messaging among agents models the interaction among components of the domain of the real world that is being managed.
   

Complexity Management tools are constructed using multi-agent software technology.