Complex systems · introduction
An introduction
“… all is held together by a natural though imperceptible chain which binds together things most distant and most different, I hold it equally impossible to know the parts without knowing the whole and to know the whole without knowing the parts in detail.”
What makes a system complex
A system is a group of interconnected elements. It becomes complex when its behaviour is determined by the interactions between those elements rather than in the elements themselves — when it's the connections, not the parts, that decide how the system behaves.
That is the difference between a complex and a complicated system. A complicated system - a calculator, an engine, a legal text - can have many intricate parts, but you can take it apart and understand it piece by piece to reconstruct the whole. In a complex system, this is impossible because the whole is also defined by how the parts interact.
Each element may follow simple local rules, with no single entity in control, and produce new, sophisticated collective behaviour or properties that none of them possesses on its own.
Why study Complex Systems
Complex systems are all around us, and many go unrecognized as complex for a long time. The reason is that when a complex system is near equilibrium, it can be locally approximated by a simple one, and the rules we derive for it will continue to hold until the system drifts from that equilibrium. Epidemics, markets, ecosystems, cities, traffic, and the human mind are all complex systems that can be understood through the same small set of ideas. And the science of complex systems provides a conceptual framework that allows us to study these systems across disciplines.
Common complex systems properties
No particular property alone defines complexity, but this set of features are the most common one. Some are general, others mark a particular kind of complex system. Each has a link to its source; some are also linked to the models available here, the models.
Emergence
see flocking →Whole-system behaviour and properties that the parts don't have on their own, and that only emerge when they interact in a whole.
SEP, “Emergent Properties” ↗Self-organization
see Game of Life →Organised global pattern arising from local interactions, with no external control.
Mitchell, Complexity (2009) ↗No central control
No single component is in charge; order is distributed across the interactions, not issued from a centre.
Mitchell, Complexity (2009) ↗Nonlinearity
Output isn't proportional to input — small causes can have outsized effects, and effects don't simply add up.
Ladyman, Lambert & Wiesner (2013) ↗Feedback
see the bullwhip →A component's output loops back to shape its own later input. Positive feedback amplifies; negative feedback stabilises.
Wiener, Cybernetics (1948) ↗Adaptation
In many complex systems the components adapt or react to the patterns they collectively create — “complex adaptive systems”.
Holland (Santa Fe Institute) ↗Networks & connectivity
What matters is the pattern of interactions, not the substrate; the same network can describe a body, a city, or an economy.
Barabási, Network Science ↗Far from equilibrium
Complex order is usually maintained by open systems driven from outside by flows of energy or matter — Prigogine's “dissipative structures”.
Prigogine, Nobel Prize (1977) ↗Sensitivity to initial conditions
see chaos →In chaotic systems, tiny differences in starting state diverge exponentially, so long-range prediction becomes practically impossible.
Lorenz (1963) ↗Criticality & power laws
see the sandpile →Poised at a critical point, a system shows events of every size — avalanches with no characteristic scale, distributed as power laws.
Bak, Tang & Wiesenfeld (1987) ↗Robustness & fragility
The system's order persists under many perturbations — yet the same structure can carry hidden, sudden modes of collapse.
Ladyman, Lambert & Wiesner (2013) ↗Hierarchy
Nested levels of organisation, where each level has concepts and regularities that aren't present one level down.
Anderson, “More Is Different” (1972) ↗