Future Enterprise- Network Science

Network science will be a critical enabler of advanced enterprise management in the 21st century.

Major advances are already being made in applying the principles of network science to social, technological and business systems and it will be vital for the future enterprise to weave sophisticated network optimisation principles into all aspects of its operations.

Network science essentially involves analysing and managing the properties and dynamics of interconnected complex systems such as social groups, the Web, power grids, supply chains, markets, ecosystems and the brain.

Such networked systems are based largely on scale-free topologies. This is the natural architecture most relevant to the world around us and is modelled on structures with a relatively small number of hubs or nodes, each with a large number of connections and a much larger number of nodes with a relatively small number of links- broadly obeying a mathematical power law.

Knowledge of network topology and dynamics allows for optimisation and prediction of the behaviour of complex system processes and is becoming increasingly vital in managing major business activities, via information systems that control vast numbers of interlinked transactions, resources, agents and events.

Failure in a tightly coupled network such as a power grid or market system of a single node may force the failure of other nodes, resulting in cascades of failures, eventually triggering a catastrophic breakdown of the whole system, as in the recent collapse of the global economy.

Examples of potential applications of network science principles include-

Economies and Markets- reducing the risk of global failure by ensuring economic networks are more robust; by closely monitoring market signals and adjusting the topology of nodes and links to reduce the risk of runaway feedback loops and conflicts between local interests and global efficiency.


Ecologies and Biodiversity- improving the sustainability of ecological systems in a period of global warming with the capacity to provide timely warning of species and resource collapse. The network model provides a powerful representation of ecological interactions among species and highlights global ecological interdependencies, which can then be re-modelled to manage risk.

Business and Finance- improving the capacity to make quality decisions regarding markets and product development, to avoid the future collapse of companies such as General Motors and Lehman Brothers. In these instances, poor decision-making was amplified by the systemic risk of runaway cascading financial asset dependencies, due to overloaded coupling strengths between nodes and indeterminate feedback loops in the myriad interconnected customer and supply networks.

The relevance of network science for the future enterprise is therefore threefold-

Firstly, many of the systems involved in business may be modelled in the future by scale-free networks, such as supply chain, investment, infrastructure, production and customer systems; with nodes representing suppliers, assets, products, consumers and customer groups.

Secondly, an organisation’s systems may be modelled by networks, with nodes representing process and activity decisions and the links represented by the dynamic flows of information feeding them.

Thirdly, the architecture of the enterprise itself may be viewed as a network of control flows between decision-makers and operational agents. As processes become more complex and time critical they will be increasingly automated, but the architecture- the information and decision-making structures and channels, will still need to be continuously optimised.

Future Trends

As forecast in previous posts, the enterprise of the future will be driven by networked architectures- patterns of linked decision processes- constantly morphing, reforming and adapting to a continuous flux of a changing global environment.
Today’s traditional hierarchical or even flat management models will be incapable of supporting tomorrow’s vastly more complex and competitive techno-social environment.

Such techno-social systems composed of technological layers operating within the larger social and physical environment that drives process application and development will need a more integrated, adaptive and intelligent framework for achieving sound management capability, underpinned by network science.

Most real world transportation, manufacturing, computing and power infrastructure networks will be linked and monitored by sensors and tags embedded in largely autonomous networked societies; constantly adapting to global evolutionary dynamics.

Network science algorithms will be developed to monitor and engineer optimal decision topologies, critical thresholds and non-linear outcomes. These will combine with AI technologies to manage complex enterprise operational and management processes.

These algorithms will apply adaptive defence mechanisms, often providing counterintuitive approaches to the engineering and control of complex techno-social systems. Such techniques will be based on the manipulation of key nodes, links and pathways to induce intentional network behavioural changes- mitigating for example potentially catastrophic outcomes.

This will represent the new Network Science Management Paradigm of the 21st century.

Future Enterprise- The Networked Enterprise

The enterprise of the future will be driven by a networked architecture- patterns of linked decision processes; constantly morphing, reforming and adapting to a continuously changing social and business environment.

The traditional hierarchical management model of the 20th century will be incapable of supporting the vastly more complex and competitive 21st paradigm of technological and social evolution.

Tomorrow's enterprise can be most effectively represented as a decision network model with decisions as nodes and information flows linking the relationships between them. This model represents an extremely powerful mechanism for understanding and optimising the adaptive enterprise of the 21st century- linked to but extending far beyond current simplistic process models.

Although process and object representations are a necessary and logical intermediary step in the evolution of enterprise system modeling and management, they fail to represent the underlying decision complexity of the real world and therefore fail to realise the true potential of a dynamic enterprise.

The core of the Networked Architecture will be the Decision Model, incorporating engineering methods based on decision pathways, with the capacity to dynamically route information and intelligence resources to critical decision-making agents in the enterprise.

This will not only involve the deployment of computing and information resources to adaptive decision nodes, but facilitate direct targeting of intelligence and problem solving capacity, enabling critical decision outcomes to be implemented in optimal time frames.

The latest 'Smart Planet' paradigm, in which the infrastructure and processes of the planet- whether manufacturing supply chains, electricity grids, water networks or traffic flows, are being re-engineered to optimise performance and achieve greener outcomes, will be the major driver for the networked enterprise of the future. The Smart Planet will demand that decisions be made more rigorously, efficiently, adaptively and therefore largely autonomously.

While SOAs focus on basic services, their capacity to implement complex decision processes is far from optimal. Current business intelligence and data warehouse software represents a halfway house towards this goal. But predictive techniques utilising AI will be the next stage, layered on current data mining and pattern recognition software and supported by a new generation of network-oriented database management systems.

Although the more far-sighted businesses are becoming aware of the need for such flexible small world network linkages, the support provided by today’s rigid organisational management architectures and philosophies has lagged well behind.

Tomorrow’s enterprise management must be far more pro-active and sentient in relation to environmental and structural change, avoiding being caught passively flat-footed in a bewildering flux of global evolution and competitive pressures.

Future Enterprise- The Intelligent Enterprise

The enterprise of the future will increasingly depend on a wide range of rigorous artificial intelligence systems, algorithms and techniques to facilitate its operation at all levels of e-commerce management.

As described in The Adaptable Enterprise blog, major decisions incorporating sophisticated levels of intelligent problem-solving will be increasingly applied autonomously within real time constraints to achieve the level of adaptability required to survive in an ever changing and uncertain global environment.

In addition web services will draw on the advances already made by the semantic web combined with the intelligent web 4.0.

A number of artificial techniques and algorithms are rapidly reaching maturity and will be an essential component of Intelligent Enterprise Architecture of the future.

Current techniques include-

Genetic algorithms- achieve solution discovery and optimisation modelled on the evolutionary natural selection process- based on the genetic operators of cross over, replication and mutation and measured against a 'fitness function'.
This technique is widely applied to solve complex design and optimisation problems.

Bayesian networks- graphical models representing multivariate probability networks- providing inference and learning based on cumulative evidence- widely used in medical diagnosis

Fuzzy Logic- based on natural non-binary methods of decision-making- assigns a

Swarm Intelligence- combines multiple cooperating components to achieve group intelligent behaviour.
Neural networks- pattern discrimination techniques modelled on neuron connection.Allows information inputs to be weighted and an activation threshold established.
Expert Systems- rule based inference techniques targeted at specific problem areas.
Intelligent Agents- designed to be adaptive to the web's dynamic environment- an agent is designed to perform a goal and learn by experience- can also act collaboratively in groups achieving higher levels of intelligence and capable of making increasingly complex decisions autonomously.

Future Trends

The above techniques will continue to be enhanced and packaged in different combinations to provide immensely powerful problem solving capability over time. The technology is slowly being applied discretely within business intelligence, data mining and planning functions of enterprise systems.

However AI is yet to realise its full potential within the enterprise by being applied to decision-making in a targeted autonomous fashion.
When this happens over the next decade, the quality of decision-making and concommitant reduction in operational and management risk is likely to be significantly improved.