A complex systems methodology to transition management

DIMETIC, Maastricht, 15-19 Oct 2007

Koen Frenken (k.frenken@geo.uu.nl)

Structure of the talk

• Some remarks on NK-models

• The power of decomposition and the example of the Wright Brothers inventing the airplane

• A complex systems methodology to transition management with an application to future sustainable car technologies

Properties of NK

• N stands for the number of components in a system

• K stands for the number of interdependencies between components

• The number of possible strings, called design space or state space or possibility space: S = 2N

• The number of local optima for K is maximum can be derived analytically # = 2N/(1+ N)

• The fitness of local optima tends towards the mean for increasing K and N (complexity catastrophe)

• Fitness of the one global optimum increases for increasing K and N

Some thoughts

• Henderson/Clark classification- Incremental- Modular- Architectural- Radical

• Search strategies- Design space search- Function space search

• Decomposability and search[Next slides]

K. Frenken (2006) Innovation, Evolution and Complexity Theory (Edward Elgar)

Finding the global optimum

• Given that local optima exists, finding the global optimum generally requires ‘exhaustive search’ involving 2N trials

• Except for decomposable systems• If complexity refers to problem-solving difficulty,

K is not always a reliable complexity indicator

Testing the glider subsystem

Testing the system as a whole

‘Doing it Wright’

A complex systems methodology to transition management

Remainder is based on joint work with:

Malte Schwoon (Max Planck Institute Hamburg)Floortje Alkemade (Utrecht University)Marko Hekkert (Utrecht University)

Paper available at DRUID summer conference 2007 and via DIMETIC website

Objective

• To develop an empirically-based policy framework for technology assessment

• that takes into account path dependence in the design of complex technologies

Technological transition

“A technological transition is the substitution of a complex technological system by an alternative system”

Complex systems theory

• Technological systems contain many interdependent subsystems

• Changes in one part of the system create unexpected effects in other parts

• System evolution is path dependent in that a choice at one moment in time affects the likelihood of subsequent choices

Flexibility of initial step

• Design flexibility (how many optima can be reached after the initial step)

• Path flexibility (how many routes exist towards an optimum)

• Time flexibility (how many transition steps are involved in the transition)

Assumptions

• Local search

• Up-hill moves only

Properties

• 7×2×7×9×3 = 2646 possible designs

• Majority of designs has lower fitness than current system

• Many neighbouring designs have similar fitness values creating plateaus in the fitness landscape

Conclusions

• Complex patterns of mapping can be traced empirically by collecting comprehensive data on design space and fitness values

• Choice of subsystem innovation affects flexibility in many respects

• Applicable to many technologies

• Assumptions about search behaviour and number of agents should be qualified