The goal of self-organization concept is to bring order from an existing state to a more desired state. Self-organization in computer systems has several interpretations - as many as the systems built and contexts used. An often used interpretation is configuring systems to achieve optimal performance in the desired state. As the world of computer systems is expanding rapidly towards pervasive, complex, digital business ecosystems, there is a necessity to interpret what self-organization means for the several layers of computer systems (vertical) which in turn form networks of (horizontal and vertical, complex) systems. I will attempt to highlight the use of just two terms - uncertainty and feedback in the context of self-organization of such complex systems as these terms are often associated with self-organization. Several interpretations, in turn, come into play with this terminology for different layers ranging from the network, infrastructure layers to the user layer of computer systems. The varied interpretations with just the two terms associated with self-organization for ecosystems suggest that building digital business ecosystems combining traditional settings with new digital business models require clarity and attention to detail.
Uncertainty is experienced with surprise and little information revealed on the disorder. The distribution of patterns, timeliness, and complexity of the associated variables are some of the causes of the uncertainty affecting the desired state of the system. Designing a matching system whether it is for renters and owners, or of males and females for a dating event can result in quite uncertain outcomes depending on the several associated factors. More difficult factors are how to get people on board when there are no apparent incentives or pull to draw them to use the system. Once you get people on board, what about handling the details of the actors involved? The cause of uncertainty for the matching system could be all the factors or just a factor for the success or failure of the event. The key to resolving uncertainties is to turn to clear assumptions on what can go wrong. Statistical reasoning can be the rescue here. Again, just using the best available statistical tool and a data scientist to interpret the data with the tool is not sufficient. Expert knowledge to understand and predict the nature of uncertainty with the model of distribution patterns is the first step. This needs to be contextually combined with the feedback data. Both the positive and negative implications need to be well thought of.
Some ecosystems or networks are so complex (a city is a very complex network with several computer systems) that questions arise if they can be self-organized. Going deeper into the mathematical reasoning for modeling city congestion, there are mathematical functions like Lyapunav functions to prove that it is possible. "A Lyapunav function maps configuration at time t into the real numbers and satisfies two assumptions. First, if the transition function is not at equilibrium, the value of the Lyapunav function falls by a fixed amount. And second, the Lyapunav function has a minimum value. If both assumptions hold. then the dynamical system must attain an equilibrium" - Page (see References on Model Thinker below). Page discusses how cities self-organize with a "city activities model" in a very clear and readable way. Each individual or a corner shop does contribute to the overall congestions. Adjusting their schedules to meet less traffic results in reaching equilibrium for themselves and reduce the overall city congestion.
Whether the context is a modestly big network like a supply chain network or understanding the effect of a person’s portfolio of modest investments, feedback mechanisms help analyze the system dynamics that come into play. Here, the key is to pay attention to feedback from the system - positive, negative, combined effects of positive and negative resulting predator-prey scenarios. These feedbacks are challenging not just because the scale of the network growth brings in threats from competing forces. The challenge exists even with the combined direct and indirect effects of a couple of positive and a single negative feedbacks as they are beyond human reasoning to predict. Such effects reveal with constructing flows with boxes and arrows and developing models applying mathematical reasoning.
Whether the context is a modestly big network like a supply chain network or understanding the effect of a person’s portfolio of modest investments, feedback mechanisms help analyze the system dynamics that come into play. Here, the key is to pay attention to feedback from the system - positive, negative, combined effects of positive and negative resulting predator-prey scenarios. These feedbacks are challenging not just because the scale of the network growth brings in threats from competing forces. The challenge exists even with the combined direct and indirect effects of a couple of positive and a single negative feedbacks as they are beyond human reasoning to predict. Such effects reveal with constructing flows with boxes and arrows and developing models applying mathematical reasoning.
So, how is self-organization defined and designed for the growing number of business platforms and ecosystems? How is it achieved? Take the context, steps, and models needed to achieve an order of the various layers of the system seriously. The space covered by the internet based complex adaptive system requires expertise at various layers and levels - starting from a unique market creation to identifying hundreds of variables affecting these various layers and continuous experimentation with the effects. It is already becoming clear to practitioners and researchers in the field of platforms that nature-inspired patterns for self-organization and self-adaptation can be the source for effective digital ecosystem designs. And that may just be the beginning.
The limitations of current software engineering (inspired by systems engineering) approach for self-organization with configuration management and feedback control loops are that these mechanisms work well in systems whose size and scope is suitable for developers (human intervention) to manage. Web-based business ecosystems are interconnected firms with abilities to co-create, collaborate and possibly no specific hierarchical management and much more. Ecosystems are clearly out of scope for the traditional methods of software configuration used in software engineering practices for information systems either for predictability or managing large scale effects of uncertainty. A few of the often mentioned topics on platforms or ecosystems use nature-inspired protocols for attaining an equilibrium, dealing with competition or monitoring behaviors and creating patterns of organization.
The context for swarm effects is clear in the mass movements with people coming together naturally in a specific location or around an idea for a specific purpose. These effects are observed in nature in bees or other insects. This idea is put to use in developing new transportation ecosystems as relevant as creating a macro-situation map for routing drivers with situational awareness. Tracking human behaviors in an organization in response to a transformational change is a related context.
Viral growth effects are often associated as mechanisms to cause desired marketing effects in ecosystems. The design of the protocols causing the viral growth effects is inspired by nature - how gossip spreads or epidemics spread in a very open, adaptable, and creative way from host to host forming a network with the gossiping people thriving on topics that interest them or spreading the disease.
Leaving environmental trails or traces to gather information on the go and form patterns to create a massive, collectively co-ordinated effect is the mechanism of stigmergy. Stigmergy is demonstrated by ants, termites (social insects), bees, forming colonies, birds flocking and flying together. This phenomenon is evident in Multi-agent model simulations in video games
The challenge is to come up with mechanisms paying attention to the deeper implications of the organic approaches taken by firms for the life cycles of the member firms of the ecosystems. Perhaps nature's models of self-organized communities which allow for co-existence in nature's ecosystems offer better solutions. Understanding the models behind the phenomenon observed in the markets of platforms and ecosystems provide better reasoning and power to analyze, plan or predict for action.
Leaving environmental trails or traces to gather information on the go and form patterns to create a massive, collectively co-ordinated effect is the mechanism of stigmergy. Stigmergy is demonstrated by ants, termites (social insects), bees, forming colonies, birds flocking and flying together. This phenomenon is evident in Multi-agent model simulations in video games
The challenge is to come up with mechanisms paying attention to the deeper implications of the organic approaches taken by firms for the life cycles of the member firms of the ecosystems. Perhaps nature's models of self-organized communities which allow for co-existence in nature's ecosystems offer better solutions. Understanding the models behind the phenomenon observed in the markets of platforms and ecosystems provide better reasoning and power to analyze, plan or predict for action.
References:
Classical Papers - Principles of the self-organizing system
E:CO Special Double Issue Vol. 6 Nos. 1-2 2004 pp. 102-126
Industries, Ecosystems, Platforms, and Architectures: Rethinking our Strategy Constructs at the Aggregate Level
https://www.researchgate.net/publication/291372335_Ecosystem_Platform_and_Industry_Architecture_Research_Re-focusing_the_Agenda
Key Affordances of Platform-as-a-Service: Self-Organization and Continuous Feedback (Researchgate)
https://www.cio.com/article/3193073/artificial-intelligence/why-is-stigmergy-a-good-platform-for-swarm-intelligence.html
https://www.huffingtonpost.com/entry/understanding-cryptocurrencies-with-stigmergy_us_5a387452e4b0578d1beb7243
https://www.huffingtonpost.com/entry/understanding-cryptocurrencies-with-stigmergy_us_5a387452e4b0578d1beb7243
The Model thinker by Scott E. Page
Chapter 12: Entropy - Modeling uncertainty
Chapter 18: Systems Dynamics Models
Chapter 19: Threshold models with feedbacks
Chapter 26: Models of learning
Chapter 16: Lyapunov Functions and Equilibrium
(Self-organization - New York and Disney world).
https://www.ineteconomics.org/uploads/papers/WP_92-Frydman-et-al-KUH.pdf
https://www.ineteconomics.org/uploads/papers/WP_92-Frydman-et-al-KUH.pdf
