Collective Adaptive Systems (CAS)

The BPM 2014 conference, Sept 7-12, has been moved from Israel to Eindhoven Holland (because of unrest in the middle east) and I will be giving a keynote on Wednesday Sept 10.  There will be an interesting workshop on Business Processes in Collective Adaptive Systems (BPCAS’14) on Monday, associated with a group called FoCAS (Fundamentals of Collective Adaptive Systems).

What is a Collective Adaptive System?

Also sometimes called “Adaptive Collective Systems,” they are described as “heterogeneous collections of autonomous task-oriented systems
that cooperate on common goals forming a collective system.”  While being wide open to interpretation, there is a key point that the units are assumed to (potentially) autonomous.  I think this is a more natural way of looking at human organizations which form automatically from humans who are themselves quite complex and autonomous.

FoCAS cropped-focas-web-logo2describes its purpose:  “The socio-technical fabric of our society more and more depends on systems that are constructed as a collective of heterogeneous components and that are tightly entangled with humans and social structures. Their components increasingly need to be able to evolve, collaborate and function as a part of an artificial society.

Nature – A strong orientation to working the way that biological systems work.  Natural systems are referenced frequently as they try to tease out the essential capabilities behind the working of ecosystems, cellular systems, herd dynamics, etc.  I particularly like the non-machine, non-Taylorist approach.

Automated or Facilitated? – There are a mix of approaches.  Some of the research seems oriented toward facilitating humans in an organization, and some is toward replacing humans with automated, yet flexible, systems.

Non-Uniform – Another thing I like about this approach is they do not assume that there is a single uniform process system.  So much of BPM research assumes that all actors will interact with a single process.  This approach assumes from the beginning that there will be many diverse components interacting in complex ways.  Diversity is the important ingredient for stability in the face of unexpected changes.

FoCAS offers a free book to get an overview of the situation:  “Adaptive Collective Systems: Herding Black Sheep”  offers 75 pages that cover the need and various approaches that they are trying.

Research Projects

These projects (all in Europe) are associated with FoCAS:

  • Allow Ensembles – human oriented pervasive business processes.  Define processes as flow, but it is expected in real life that the process will need to be changed.  No single system, but the idea that there will be a large number of separate systems.  There are different goals at different levels: individual and collective.  Non-functional requirements are called “utilities” (e.g. reduce smog, increase efficiency).  Processes are defined in cells and cells collaborate.  Example given is a travel scenario for two people that has to be adapted.  Clearly the person involved is able to modify the route, although it is not clear whether they want to make this ‘automatic’ or not.  Supply chain is another example.
  • Assisi | bf – Project to interact with collections of animals (or presumably humans) in order to influence behavior.  Examples were bees and fish.  Ultimately this is for influencing huma n”swarm intelligence.”  Compares their work to google, wikipedia, facebook, and twitter.
  • CASSTING – Stands for collective adaptive systems (CAS) and Synthesis With Non-Zero-Sum Games (STING).  They use a game theory approach to evolve the correct independent units.
  • DIVERSIFY – Goal is to learn how biodiversity emerges in ecosystems. These systems are plastic and able to adapt to many kinds of changes.  This is quite different from the software we use today which is usually picked from one a a small number of varients.  This is fragile.  If systems can be made more diverse, it might be more robust.  One challenge is that it is a overlap between math/statistics, computer science, and biology.  What is the nature of software that support diversity.  Simply scrambling code will not work.
  • QUANTICOL – Quantitative modeling of collective adaptive systems.  Made of components which have state and communicate with other components.  Looking at smart grid. Edinborough has a bus system which reports positions every 30 seconds.  They are looking at how to adapt to emerging roadwork or traffic patterns.  Can traffic lights be tweaked to optimize the system.   Looks to me like ‘automatic’ adapting without explicit ways for people to manipulate the system.
  • Smart Society – Key to making something robust is diversity. Ethics, trust, and reputation.  A large semantic gap between human systems and computer systems.
  • Swarm organ – Machines and technology are quite fragile.  Biology can do amazing things, like self healing.  Studying morphgenesis: how do cells form organs, and there are multiple strategies for how this might happen.  Why do it one way, and why do it another.  The idea is that you might make self-organizing systems that form themselves into the systems that we use.


Robust systems will need to be designed in this way, with a lot of collaborating yet diverse systems each advocating different goals.  People need to be part of these systems, and must interact fluidly with them.  The collective adaptive system approach a distinctly non-Taylorist approach that is worth watching.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s