Download as pdf or txt
Download as pdf or txt
You are on page 1of 58

Why design

thinking works?

Or
Design Thinking Unpacked: An
Evolutionary Algorithm
J. M. Korhonen & L. Hassi

in this presentation
- why design thinking works
- when does it work
- what does it mean in practice

Design thinking-like approach


in practice is defined here as:
- multidisciplinary teams
- human-centred exploration
- fast and iterative prototyping
(process perspective; Jahnke 2009)

what designers do

Design is defined here as a


knowledge-generating activity

Product development =

Product development =
search for best possible designs

Imagine an (almost) infinite


library of all designs...

(cf. The Library of Babel by Jorge


Luis Borges)

Trinity College, Dublin

If we visualize whats in the


library:
(mobile phones section)

Differences in design

Differences in design

Differences in design

Differences in design

Utility (fitness for purpose)

Differences in design

EXAMPLE CASE: Jannes choice, 2004


Utility (fitness for purpose)
Differences in design

EXAMPLE CASE: Jannes choice, 2004


Utility (fitness for purpose)

Differences in design

Utility (fitness for purpose)

Differences in design

Utility (fitness for purpose)

Differences in design

FITNESS LANDSCAPE
Utility (fitness for purpose)
Differences in design

PERFECTLY ORDERED (NON-RANDOM)


Utility (fitness for purpose)

Problem type: Defined, quantitative


Differences in design

ROUGH-CORRELATED (REAL LIFE)


Utility (fitness for purpose)

Problem type: Wicked, qualitative


Differences in design

What does rough-correlated


fitness landscape mean in
practice?

Usually, small changes have


small effects on fitness for
purpose...

Mirra Chair (c) Herman Miller

But sometimes, small changes


can have large effects on
fitness...

Mirra Chair (c) Herman Miller

Mirra Chair (c) Herman Miller

?? ?

Mirra Chair (c) Herman Miller

[x] Metric
[x] Imperial

On the other hand, some large


changes may have only small
effects on the fitness for
purpose...

Mirra Chair (c) Herman Miller, Office Chair (c) vcf.com

ROUGH-CORRELATED (REAL LIFE)


Utility (fitness for purpose)
Differences in design

How to reach the highest


possible peaks?

The optimum strategy for getting


to the top in rough-correlated
landscapes:
evolutionary algorithms

Informal definition:
Algorithm is a process that
performs some sequence of
operations

EVOLUTIONARY ALGORITHM
Utility (fitness for purpose)

Differences in design

EVOLUTIONARY ALGORITHM
Utility (fitness for purpose)

X
X

Differences in design

EVOLUTIONARY ALGORITHM
Utility (fitness for purpose)

X
X

Differences in design

evolutionary algorithm =
- radical experimentation
- incremental improvement
- test, eliminate, retain

evolutionary algorithm =
- diversity of ideas
- iterative prototyping
- rapid real-life testing

evolutionary algorithm...?
- multidisciplinary teams
- human-centred exploration
- fast and iterative prototyping
evolutionary algorithm...?

Evolutionary algorithm

Design thinking

Radical experimentation (lots of ideas)

Multidisciplinary teams

Incremental
improvement

Human-centred
exploration

Test, eliminate, retain

Fast and iterative


prototyping

Some implications:
- Explaining design thinking
- When to use design thinking
(- NPD process modeling)
(- Technology S-curves)

Provisional theoretical
explanation: why design
thinking works

Provisional theoretical
explanation: why design
thinking works
(and where it works best)

In short, design thinking-like


approaches may be
theoretically near-optimum
strategies when the fitness
landscape is rough-correlated

In short, design thinking-like


approaches may be
theoretically near-optimum
strategies when the fitness
landscape is rough-correlated
(that is, in most cases)

Could we estimate the proper


exploratory/exploitative
(inductive/deductive) mix in
actual projects?

Could we estimate the proper


exploratory/exploitative
(inductive/deductive) mix in
actual projects?
Could this affect resource
planning?

When to use design thinking

PERFECTLY ORDERED (NON-RANDOM)


Utility (fitness for purpose)

NOT GOOD
Problem type: Defined, quantitative
Differences in design

ROUGH-CORRELATED (REAL LIFE)


Utility (fitness for purpose)

GOOD
Problem type: Wicked, qualitative
Differences in design

HOWEVER, when zooming in


by defining the problem better,
qualitative can become quantitative

Problem type: Defined, quantitative

Well-defined problems are best


solved through formal,
analytical approaches

...of course, getting to welldefined is the trick: engineers


are really good at finding
answers, but how to ask the
questions?

FI(n)NISH

janne.korhonen@seos.fi
lotta.hassi@tkk.fi
www.seos.fi
www.aaltodesignfactory.fi

You might also like