(1) the complete title of one (or more) paper(s) published in the open
literature describing the work that the author claims describes a
human-competitive result,
Deb, K. and Srinivasan, A. (in press). Innovization: Innovating design
principles through optimization. Proceedings of the Genetic and Evolutionary
Computation Conference (GECCO-2006), to be held in Seattle, USA during
8-12 July 2006.
(2) the name, complete physical mailing address, e-mail address, and
phone number of EACH author of EACH paper,
Kalyanmoy Deb
Professor
Department of Mechanical Engineering
Indian Institute of Technology Kanpur
Kanpur. PIN 208016, India
Email: deb@iitk.ac.in
Tel: 0091 512 259 7205 (office), 8310 (home)
Aravind Srinivasan
Graduate Student
Department of Mechanical Engineering
Indian Institute of Technology Kanpur
Kanpur. PIN 208016, India
Email: aravinds@iitk.ac.in
Tel: 0091 512 259 7668 (office)
(3) the name of the corresponding author (i.e., the author to whom
notices will be sent concerning the competition),
Kalyanmoy Deb (deb@iitk.ac.in)
(4) the abstract of the paper(s),
Designers and practitioners routinely look for new and improved
solutions and often resort to forming and solving an optimization
problem in order to find the best solution corresponding to a chosen
objective or a design goal. Although such a single optimum solution
may bring out some innovation in its design or working principle,
often they could be very specific to the chosen objective.
In this paper, we suggest a multi-objective optimization strategy
which is capable of finding more than one optimal solution, each
corresponding to a certain trade-off among the objectives. Thereafter,
we suggest a systematic information retrieval strategy to decipher
salient design principles which are common to these multiple optimal
solutions. Such a dual task (we call it an 'innovization' task)
is demonstrated to discover useful and importantly, innovative,
relationships among objectives and decision variables which must be
present in solution to make it a high-performing, optimal solution.
Such valuable information about a design problem are difficult to
achieve by any other scientific method. In addition to
simply finding an optimal solution, the innovization task allows one
to reveal important 'recipes' and 'blue-prints' for a solving
a problem optimally, which are not often intuitive and beyond human
comprehension. In a number of engineering component design problems,
we discover useful relationships, such as a fixed thickness and
a fixed difference in inner and outer diameters of a multi-disk brake
system are properties of all optimal brakes ranging from small to
large stopping-time considerations (with a trade-off with brake
mass). Similarly, in a spring design
problem, small to large sized (having large to small developed stress)
optimal springs having different
wire diameters, spring diameters, and number of turns, all having a
specific spring stiffness qualify to be the optimal designs.
Such information are not often human-conceivable and intuitive. The
proposed innovization procedure allows a way to unveil such useful
design principles common to multiple optimal solutions.
(5) a list containing one or more of the eight letters (A, B, C, D, E,
F, G, or H) that correspond to the criteria (see above) that the author
claims that the work satisfies,
We feel that our work falls into the following two categories:
D) The result is publishable in its own right as a new scientific
result - independent of the fact that the result was mechanically
created.
(G) The result solves a problem of indisputable difficulty in its field.
(6) a statement stating why the result satisfies the criteria that the
contestant claims (see the examples below as a guide to aid in
constructing this part of the submission),
We now argue why we think our entry satisfies each of the above two
categories.
D) The innovization principles obtained by the post-optimality analysis
are mathematical relationships among decision variables and
objectives, which are derived from multiple optima.
In the context of an electric motor design
problem, the deciphered innovization principles may translate to the
following concept. An optimal procedure of designing a motor for increased
power output is to proportionately increase the armature diameter
and have a fixed wire diameter. After such relationships are
deciphered from the multiple trade-off optimal solutions (say,
simultaneous maximization of power output and minimization of size of
motor, in the above example), they can be substituted in the original
optimization problem formulation to simply the problem and optimality
of the relationships can be tested using mathematical optimization
principles. Since the relationships are properties of trade-off
*optimal* solutions, they are difficult to find by other means, except
by first finding a representative set of optimal solutions and then
looking for interesting relationships among them.
Thus, although the results are mechanically created by using an
evolutionary multi-objective optimization procedure followed by
a number of local searches and other multi-objective optimization aids
to increase the confidence in the optimality of obtained solutions,
the information retrieval strategy performed by authors in
specific engineering design problems (the idea is also followed by other
researchers) has always discovered innovative and
useful design principles which were publishable in domain specific
journals.
G) As mentioned above, the innovization task is also a unique
procedure of obtaining innovative design principles in a
problem. The innovization idea is new and systematic,
and there does not exist any known methodology for achieving a similar
task. It is worth mentioning here that
the 'Monotonicity Analysis' of identifying optimal solutions without
actually performing an optimization task can find some such useful
relationships among decision variables, but the technique is only
applicable to monotonic objective functions and constraints and
is not at all a generic approach.
(7) a full citation of the paper (that is, author names; publication
date; name of journal, conference, technical report, thesis, book, or
book chapter; name of editors, if applicable, of the journal or edited
book; publisher name; publisher city; page numbers, if applicable);
Deb, K. and Srinivasan, A. (in press). Innovization: Innovating design
principles through optimization. Proceedings of the Genetic and Evolutionary
Computation Conference (GECCO-2006), to be held in Seattle, USA during
8-12 July 2006.
(8) a statement either that "any prize money, if any, is to be divided
equally among the co-authors" OR a specific percentage breakdown as to
how the prize money, if any, is to be divided among the co-authors; and
80% (Deb) - 20% (Aravind)
(9) a statement stating why the judges should consider the entry as
"best" in comparison to other entries that may also be
"human-competitive."
The idea proposed in our entry is a level higher than probably the most
human-competitive results proposed so far. It is likely
the most other results are based on a single optimal solution
corresponding to minimizing or maximizing a specific objective
function related to the problem. Our innovization procedure would
allow multiple conflicting objectives to be considered, thereby
allowing to find a number of optimal solutions, trading-off among
conflicting objectives. Moreover, instead of analyzing human
competitiveness and innovation associated with a single optimal
solution, our innovization procedure brings out common principles
associated with a number of such trade-off yet optimal solutions (including
the single-objective optimum solution), which would remain as
higher-level features of human competitiveness and innovation. In this
entry, we suggest a systematic procedure of performing this task and
demonstrate its usage on a number of engineering design case
studies. In each case, previously unknown yet interesting human
competitive design principles were unveiled. In some cases, human
conceivable results are also found.
The most unique aspect of our entry is that the proposed innovization
technique can be used to *learn* about salient properties of optimal
(high-performing) solutions, trading-off two or more goals of
design. To put it in the context of a concrete example, the proposed
technique is way to unveil how different motoring parameters, such as
coil diameter, number of turns, armature diameter, etc., must be varied with
respect to each other in designing minimum-sized motors with an increasing
delivered power. The idea goes beyond simply finding an optimal
design, but trying to understand how to make and what makes a
design optimal.