TIES598 Nonlinear Multiobjective Optimizationusers.jyu.fi/~jhaka/ties598/TIES598_visualization.pdf · Interactive decision maps –A. V. Lotov et al., Interactive Decision Maps: Approximation

TIES598 Nonlinear Multiobjective Optimization

Visualization aspects in multiobjective optimization

spring 2017

Jussi Hakanen & Vesa [email protected]

Contents

Visualization as a field of science

Basic visualization for 2 & 3 objectives

Visualization techniques for more than 3 objectives

Existing visualizations used in MOO

Advanced technologies

Teaser

https://www.youtube.com/watch?v=5tfnmhl-A54

Visualization research fieldsScientific visualization

– Visualization of three-dimensional phenomena (architectural, meteorological, medical, biological, etc.), emphasis being on realistic renderings of volumes, surfaces, illumination sources etc., perhaps with a dynamic (time) component

Information visualization– Information visualization is the use of computer-supported, interactive,

visual representations of abstract data to amplify cognition*

Visual analytics (http://www.visual-analytics.eu/) – Science of analytical reasoning facilitated by interactive visual

interfaces**– Visual analytics combines automated analysis techniques with

interactive visualizations for an effective understanding, reasoning and decision making on the basis of very large and complex data sets***

– https://www.youtube.com/watch?v=5i3xbitEVfs

* S. Card, J. Mackinlay, and B. Shneiderman. Readings in Information Visualization: Using Vision to Think.

Morgan Stanley Publishers, 1999

** J. J. Thomas & K. A. Cook, A visual analytics agenda, IEEE Computer Graphics and Applications, 26, 2006

*** D. Keim et al., Visual analytics: Definition, process, and challenges. In: Information Visualization: Human-

Centered Issues and Perspectives, 2008

Visualization in multiobjective optimization

Typically, solutions in the objective space are visualized– Objective space: performance of solutions– Decision space: implementation of solutions

Some examples of visualizations combining both spacesRecent survey– K. Miettinen, Survey of methods to visualize alternatives in

multiple criteria decision making problems, OR Spectrum, 36, 2014

Visualization for more than 3 objectives: indirect visualization has to be used

Visualizations for 2D/3D

Direct visualization of

– Pareto front

– Individual (nondominated solutions)

Scatter plot

Dimensionality reduction

Dimensionality of the objective space can also be reduced to enable visualization, e.g.

– Principal component analysis (PCA)

– Self organizing maps (SOM): nonlinear generalization of PCA

– Multidimensional scaling (MDS): maps the solutions on a plane while trying to preserve distances between them

Visualizations for more than threeobjectives

Indirect visualization

Typically a small set of individual solutions are visualized

Parallel coordinate plots

• Web-based PCP tool: https://reed.cee.cornell.edu/parallel-axis-categories/parallel/

• Interesting blog:

https://waterprogramming.wordpress.com/?s=visualization&submit=Search

Other visualization techniques…

Existing visualizations used in MOO

knowCube– H. L. Trinkaus & T. Hanne, knowCube: A visual and interactive support for

multicriteria decision making, Computers and Operations Research, 32, 2005

Existing visualizations used in MOO

Interactive decision maps– A. V. Lotov et al., Interactive Decision Maps: Approximation and

Visualization of Pareto Frontier, Kluwer Academic Publishers, 2004

Existing visualizations used in MOO

Heat map visualization– J. Hettenhausen et al., Interactive multi-objective particle swarm optimization with

heatmap-visualization-based user interface, Engineering Optimization, 42, 2010

Existing visualizations used in MOO

RadVis: maps M-dimensional points to 2D space using nonlinear mapping

3D radial coordinate visualization (3D RadVis)– A. Ibrahim et al., 3D-RadVis: Visualization of Pareto front in many-objective optimization, In: 2016 IEEE CEC

conference, 2016

3D RadVis: third dimension for the shape and convergence of the solution set (distance from a reference hyperplane)

More illustrations in: http://ieeexplore.ieee.org/document/7743865/

A demo

IND-NIMBUS

Optimus

Recommendations for user interfaces

In the visualization fields

Use linked visualizations where different types of visualizations are connected such that a solution highlighted in any visualization becomes also highlighted in other visualizations

Utilize interactive visualization techniques where the user can manipulate them making the system feel responsive

How this could look like…?

f1

f4

f3

f2

f5

f1 f2 f3 f4 f5

1 15.979 417.52 22.817 15026.0 9655.8

2 16.111 426.55 1.521 14444.0 8947.0

3 16.466 421.95 20.998 14976.0 9729.6

4 16.851 415.46 22.75 15003.0 9721.0

5 17.404 416.8 17.513 14971.0 9762.7

Solution 2

Solutions after 1st interaction

f1 f2 f3 f4 f5

1 15.979 417.52 22.817 15026.0 9655.8

2 16.111 426.55 1.521 14444.0 8947.0

3 16.466 421.95 20.998 14976.0 9729.6

4 16.851 415.46 22.75 15003.0 9721.0

5 17.404 416.8 17.513 14971.0 9762.7

6 16.007 425.15 8.0389 14591.0 9132.4

7 16.077 425.16 10.999 14698.0 9265.4

best worstSolution 7

Solutions after 2nd interaction

Virtual reality vs. augmented reality

What is the difference?

– https://www.youtube.com/watch?v=aSjUxM5f-eM&feature=youtu.be&t=27

– https://www.youtube.com/watch?v=Eh24LpidJug

A look to the future…

https://www.youtube.com/watch?v=vg0A9Ve7SxE

Links

Firefighters: https://www.youtube.com/watch?v=5tfnmhl-A54Rolls Royce shore control: https://www.youtube.com/watch?v=vg0A9Ve7SxEElevator maintenance training: https://www.youtube.com/watch?v=8OWhGiyR4NsPlant maintenance: https://www.youtube.com/watch?v=QTuKcm8s4QQNASA: https://www.youtube.com/watch?v=IcJ-JuA_K7U&t=1m21s

VR vs. AR: https://www.youtube.com/watch?v=Eh24LpidJug, https://www.youtube.com/watch?v=aSjUxM5f-eMMixed reality: https://www.youtube.com/watch?v=Ic_M6WoRZ7kHolograph: https://www.youtube.com/watch?v=vOKVofs5rEgVisual analytics: https://www.youtube.com/watch?v=5i3xbitEVfs

Material

K. Miettinen, Survey of methods to visualize alternatives in multiple criteria decision making problems, OR Spectrum, 36:3-37, 2014T. Tusar & B. Filipic, Visualization of Pareto Front Approximations in Evolutionary Multiobjective Optimization: A Critical Review and the Prosection Method, IEEE Transactions on Evolutionary Computation, 19:225-245 , 2015A. Ibrahim et al., 3D-RadVis: Visualization of Pareto front in many-objective optimization, In: 2016 IEEE CEC, IEEE, 736-745 , 2016Z. He & G. Yen, Visualization and Performance Metric in ManyObjective Optimization, IEEE Transactions on Evolutionary Computation, 20:386-402, 2016J. Kehrer & H. Hauser, Visualization and visual analysis of multifaceted scientific data: A survey, IEEE Transactions on Visualization and Computer Graphics, 19(3):495–512, 2013S. Liu et al., A survey on information visualization: recent advances and challenges, The Visual Computer, 30(12):1373–1393, 2014S. Tarkkanen et al., Incremental user-interface development for interactive multiobjective optimization, Expert Systems with Applications, 40:3220–3232, 2013J. J. Thomas & K. A. Cook, A visual analytics agenda, IEEE Computer Graphics and Applications, 26(1):10–13, 2006