Professional Documents
Culture Documents
Christopher Lueg
School of Engineering & ICT
University of Tasmania
christopher.lueg@utas.edu.au
StretchBack app to visualize the locations on a map and to
add specific constraints such as total time and preferred
transit times. The researcher is then plots a route covering
the most important locations. They want to add two more
locations, but as their path approaches the first it gets
stretched and harder to pull. The researcher realises that
theyve nearly hit their time limit and decides to choose the
second of the two locations, they think it will result in a
more pleasant end to their trip.
The scenario points out a number of constraints that
highlight that route planning for leisurely travel is an
instance of a class of complex problems where solutions
must satisfy multiple preferences while meeting specific
constraints, such as total available time, transit times, and
budget constraints.
Unlike determining optimal routes for supply chains, just to
give an example, the criteria in leisurely travel are often
subjective in nature. Such subjective criteria include, for
example, the relative importance of points of interest (e.g.,
locations that are interesting because of their natural beauty
vs points of historical significance), how much time is
required to enjoy a location and how much transit time
between points of interest is acceptable.
We are primarily interested in leveraging human perception
and the idea of "thinking with your fingers" (see also
[1][3][5]). In some situations an analytical solution might
provide a solution faster than this approach, however this
may not be preferable as it does not allow the user to learn
or explore the decision space, or to engage with subjective
aspects of the decision making process.
In order to allow for interactivity we translate constraints
into elastic properties that can be played with to figure out a
route that meets existing constraints and preferences.
Playing with elastic properties features in the touch based
computer game World of Goo (www.worldofgoo.com) and
related work has been done in constraint programming (e.g.,
[4]) but we are not aware of applications of elastic
constraints in interactive route planning.
In what follows we discuss how the metaphor of a rubber
band is used to translate the aforementioned constraints into
an interactive, touch-based human computer interface. First
we discuss the elastic properties of rubber bands that allow
users to perceive constraints. Mapping constraints onto
these properties would then allow users to have immediate
Elastic properties
(a) Unchanged
(c) Tension
(b) Attenuation
(c) Colour
2.
3.
(b) Constraint
Recognition
Standard
Usability
All Limit
Feedback
Limit
Approached
Feedback
Limit
Reached
Feedback
freq.
pctl.
freq.
pctl.
freq.
pctl.
freq.
pctl.
100
100
100
100
100
100
86
100
100
100
71
100
98
94
57
100
95
88
43
100
44
90
71
29
88
53
49
41
14
63
(d) Visual/exploratory
decision making
Thickness
Tension
Colour
16
12
11
13
11
10
13
10
13
Thickness
Colour