Professional Documents
Culture Documents
Us 7996344
Us 7996344
4B1
START
502
Receive a description of a
product to be optimized
504
Define a design space by selecting a set of
design variables and an objective function
space with at least two design objective
functions for MOEA optimization
508
Conduct a plurality of independent MOEA based
optimizations, each optimization is configured to
have different characteristics in forms of randomly
created initial populations (i.e., design alternatives)
and/or different evolutionary algorithm
508
Çombiné Paretc. front $olutiºns
from all optimizations at a
predetermined checkpoint
510
hether overal Continue the plurality
stopping criterion of MQEA based
has been met optimizations to next
checkpoint
U.S. Patent Aug. 9, 2011 Sheet 2 of 9 US 7.996,344 B1
154
156
ls only one yes The feasible One dominates
Solution
the infeasible One
infeasible?
nC)
160
162
Are both yes The less infeasible One
Solutions
Cominates the other
infeasible?
164 nC)
Both X and Y are
feasible Solutions
172
S X not Worse than Y in al yes Solution X dominates
objectives and X strictly Solution Y
better in at least One
objective?
ITO
176
S Y not Worse than X in a?
yes Solution Y dominates
objectives and Y strictly Solution X
better in at least One
objective?
In O
178
Solution X is non-dominated with
respect to Solution Y FIG 1 C
U.S. Patent Aug. 9, 2011 Sheet 3 of 9 US 7.996,344 B1
Crowding
Distance
d = a +b.
FIG. 2
U.S. Patent Aug. 9, 2011 Sheet 4 of 9 US 7.996,344 B1
First Second
Optimization | Optimization
302a : 302b
- - - - -- - - - - -- - - -- --- - - ---- - - - -F - -- ---- -- ----- - ----- - - - ----- -
! A
Initial A
Generation
304a
: 304b
O | O
O | O
I
!
I
:
A
N" Generation A
I A
306a I
306b
I
FIG. 3A
U.S. Patent Aug. 9, 2011 Sheet 5 of 9 US 7.996,344 B1
First : Second
Optimization : Optimization
312a 3.12b
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - F – — — — — —— — — — — — — — — — — —— — — — — — —— — — —
I
I
I
Initial
Generation A
A A
314a 31.4b
O O
315a is © 315b
N" Generation
316a 316b
318 320
FIG. 3B
U.S. Patent Aug. 9, 2011 Sheet 6 of 9 US 7.996,344 B1
I
I I I
© Initial
Generation
;
I
I
:
I
I
:
I
I
I I I
I I I
| I I
| | |
O I ! O
405a e ! © 405m
I I
!
I
I I I
I I I
! !
| | |
;
I
: :!
© I I
© Nth Generation
:
I I I
ee
I I I
I I I
|I |I |!
406a l I 406m
FIG. 4
U.S. Patent Aug. 9, 2011 Sheet 7 of 9 US 7.996,344 B1
500
502
Receive a description of a
product to be optimized
504
Define a design space by selecting a set of
design variables and an objective function
space with at least two design objective
functions for MOEA optimization
506
Conduct a plurality of independent MOEA based
optimizations, each optimization is configured to
have different characteristics in forms of randomly
created initial populations (i.e., design alternatives)
and/or different evolutionary algorithm
508
Combine Pareto front Solutions
from all optimizations at a
predetermined checkpoint
END
FIG. 5
U.S. Patent Aug. 9, 2011 Sheet 8 of 9 US 7.996,344 B1
602
Create initial generation of populations (i.e.,
design alternatives) from the design space
604
Evaluate all design alternatives (e.g.,
performing FEA, CFD, etc.)
606
as predefined
stopping criterion
been met?
608
Rank design alternatives in objective
function space
610
Create next generation of design
alternatives using evolutionary schemes
(i.e., crossover, mutation, etc.)
FIG. 6
US 7,996,344 B1
3 4
BRIEF DESCRIPTION OF THE DRAWINGS herein with respect to these figures is for explanatory pur
poses as the invention extends beyond these limited embodi
These and other features, aspects, and advantages of the mentS.
present invention will be better understood with regard to the Referring first of FIG. 1A, a tubular structural member 102
following description, appended claims, and accompanying (i.e., an exemplary engineering product) is optimized in an
drawings as follows: engineering optimization with design objective of minimiz
FIG. 1A is a diagram showing a tubular member (an exem ing the weight therefore minimizing the cost for a given
plary engineering product) to be optimized using thickness as material (e.g., regular strength steel) under certain design
design variable; loading condition. It is evident that thinner thickness 104
FIG. 1B is a diagram showing exemplary solutions includ 10 would lead to a less weight structure. However, at certain
ing Pareto optimal solutions in a multi-objective function point, the structural would become too weak to standa design
space; load (e.g., structural failure due to material yielding and/or
FIG. 1C is a flowchart illustrating an exemplary process of buckling). Hence, the engineering optimization of this tubular
structure requires another design objective of maximizing the
determining non-dominated solution criterion used in imple 15 strength, which leads to a safer structure. In this exemplary
menting one embodiment of the present invention; case, thickness 104 is a design variable, which may have a
FIG. 2 is a diagram graphically showing definition of range (e.g., from one eighthofinch to half an inch) as a design
crowding distance of a solution in a Pareto front; space. Any design alternatives are selected from this space. In
FIGS. 3A-3B are diagrams showing examples of combin multi-objective evolutionary algorithm, population or design
ing solutions from two independently conducted MOEA 20 alternatives at each generation are selected from the design
based optimization, according to an embodiment of the space.
present invention; The design space is one-dimensional (e.g., a line) when
FIG. 4 is a diagram showing an example of combining a there is only one design variable. The design space becomes
plurality of MOEA based optimization to form a better con a two-dimensional area for two variables, and so on. For more
verged and diversified set of solutions; 25 than three design variables, the design space is a hyperspace
FIG. 5 is a flowchart illustrating an exemplary process of that is not possible to illustrate.
obtaining a set of better converged and diversified solutions Based on two conflicting design objectives, FIG. 1B is an
by combining Pareto optimal solutions resulted from a plu X-Y diagram showing a results of an exemplary design opti
rality of independently conducted MOEA based engineering mization. Two axes represent two different objectives in form
design optimizations, according to an embodiment of the 30 of functions f, and f. In a multi-objective optimization, there
present invention; is no one optimized solution instead there is a set of solutions
FIG. 6 is a flowchart illustrating an exemplary process of that reflects tradeoffs among objectives. In order to differen
conducting engineering optimization using genetic or evolu tiate each solution, a concept called non-domination criterion
tionary algorithm; and is used for comparing solutions.
35 FIG. 1C is a flowchart illustrating an exemplary process of
FIG. 7 is a function diagram showing salient components determining the non-domination criterion. Two design alter
of a computing device, in which an embodiment of the present natives are evaluated to obtain respective solutions X and Y
invention may be implemented. according to the design objective functions (i.e., multi-objec
DETAILED DESCRIPTION
tive) in step 152. Solution X dominates Y (step 172), if any of
40 the following three conditions is true.
1. X is feasible and Y is infeasible. (Steps 154 and 156)
In the following description, numerous specific details are 2. Both X and Y are infeasible (step 160), but X is less
set forth in order to provide a thorough understanding of the infeasible compared to Y (step 162).
present invention. However, it will become obvious to those 3. When both X and Y are feasible (step 164), the following
skilled in the art that the present invention may be practiced 45 two conditions must be satisfied (steps 170 and 172).
without these specific details. The descriptions and represen a. X is not worse than Y in all design objectives; and
tations herein are the common means used by those experi b. X is strictly better than Y in at least one design objec
enced or skilled in the art to most effectively convey the tive.
substance of their work to others skilled in the art. In other Furthermore, one can determine if the solution Y dominates X
instances, well-known methods, procedures, components, 50 (steps 175 and 176). Finally, if neither solution dominates the
and circuitry have not been described in detail to avoid unnec other, X and Y are non-dominated to each other (step 178).
essarily obscuring aspects of the present invention. FIG. 1B shows a two-objective unconstrained minimiza
Reference herein to “one embodiment” or “an embodi tion example. Each dot represents an evaluated solution of a
ment” means that a particular feature, structure, or character design alternative (e.g., one of the populations in a given
istic described in connection with the embodiment can be 55 generation of the MOEA) within area 116. In accordance with
included in at least one embodiment of the invention. The the non-domination criterion, for each diamond 122, there is
appearances of the phrase “in one embodiment” in various at least one triangle 124 that is better than the diamond 122 in
places in the specification are not necessarily all referring to at least one objective without being inferior in other objective.
the same embodiment, norare separate oralternative embodi So all individual solutions in diamond 122 are dominated by
ments mutually exclusive of other embodiments. Further, the 60 the triangles 124. Similarly, all triangles 124 are dominated
order of blocks in process flowcharts or diagrams represent by squares 126 and squares are dominated by circles 128. No
ing one or more embodiments of the invention do not inher solution represented by triangles 124 can be said better than
ently indicate any particular order nor imply any limitations any other solution represented by triangles 124, thus they are
in the invention. non-dominated with respect to each other. All individuals
Embodiments of the present invention are discussed herein 65 represented by circles 128 are non-dominated with respect to
with reference to FIGS. 1A-7. However, those skilled in the any otherindividual hence having a best or highest rank (e.g.,
art will readily appreciate that the detailed description given rank of one). If all points represented by circles are removed
US 7,996,344 B1
5 6
from FIG. 1B, then the individuals represented by squares used. As a result, different populations are created due to
126 become non-dominated with respect to all others. There different series of random or pseudo-random numbers.
fore, squares 128 are assigned next best rank (e.g., rank of Design alternatives (populations) are then evaluated and
two), and so on. ranked in an objective function space. First and second sets
In the example shown in FIG. 1B, circles 128 represent a 5 304a–b of Pareto optimal solutions of the initial generation
set of Pareto optimal solution and the line 130 connecting all shown in FIG. 3A are for the first and second optimizations
circles 128 is called the Pareto optimal front. It is noted that 302a-b, respectively. The optimization is carried out in accor
there would generally be more than one individual or solution dance with an evolutionary algorithm (e.g., NSGA-II) shown
having the same rank. as an arrow “EA” 305 for N generations. Respective sets
One aspect of Pareto optimal solutions is referred to as 10 306a-b of Pareto optimal solutions are resulted at the end of
diversity, which is quantified with two measurements: spread the N” generation. The sets 306a-b are then combined into a
and uniformity. Spread is defined as the largest diagonal combined set of Pareto optimal solutions 308. It is visually
length of the largest hypercube formed by the set of Pareto evident that the combined set 308 has a wider spread than
optimal solutions in the N-dimensional function space. either of the sets 306a-b, hence being better converged and
Spread can also be defined as volume of the largest hyper
15 diversified. When combining the sets 306a-b, only non-domi
cube. Uniformity is defined as how uniformly the solutions nated solutions are kept (see FIG. 1C for an exemplary pro
are distributed in a set of Pareto optimal solutions. Math cess of determining dominate solution). For example, solu
ematically, uniformity A can be expressed as follows: tion 309b is dominated by solution 309a, thereby solution
3095 is not included in the final combined set 310.
20 Two optimizations 312a-b are also independently con
ducted in the example shown in FIG. 3B. In this example, the
optimizations 312a-b start with a same set of design alterna
tives (i.e., populations) as the initial generation. After evalu
ating the design objection functions and ranking, first and
25 second sets 314a–b of Pareto optimal solutions are identical.
Independent optimizations are carried out with two different
evolutionary algorithms indicated by arrow ‘EA-1’ 315a and
arrow ‘EA-2' 315b. For example, ‘EA-1’ is NSGA-II, while
where d, is defined as crowding distance of a solution in ‘EA-2 is SPEA. After N generations of evolution, two sets
function space (see graphical example shown in FIG. 2), “n” 30 316a-b of Pareto optimal solutions are resulted and combined
is the total number of solutions in the set of Pareto optimal into a combined set 318. In this example, uniformity of the
solutions. Crowding distance is defined as half of the perim combined set 318 is much better than the second set 316b and
eter of the largest hypercube around a solution that does not spread of the combined set 318 is better than the first set 316a.
encompass by any other solutions. Boundary points or solu Similar to the previous example, the final combined set 320
tions (e.g., 210a and 210n) are assigned a crowding distance 35 includes only non-dominated solutions in the set 318.
of twice the distance to the nearest neighbor. Uniformity Referring now to FIG. 4, it is shown a more general
measurement is similar to the standard deviation of the example of obtaining a better diversified set of Pareto optimal
crowding distance and hence a small value of the uniformity solutions in MOEA based engineering design optimization,
metric is desired, thereby achieving a good distribution of according to an embodiment of the present invention. A num
solutions. 40 ber of MOEA based optimizations (first to M.” 402a-m) are
FIG. 2 shows an exemplary set of Pareto optimal solutions conducted independently with arbitrary initial generation of
(shown as dots) 210a-n in a design objective function space design alternatives. Evolutions are carried out with same or
represented by two functions (i.e., a two-dimensional func different evolutionary algorithms (i.e., “EA-1’ 405a . . . ‘EA
tion space with f, and f, as the two axes). Pareto optimal front M’ 405m). M sets of Pareto solutions 406a-m are combined
is represented by dotted line 230. The distance 240 between 45 into a combined set 408 after N generations. It is noted that M
two extreme end solutions 210a and 210n is referred to as and N are positive integers or whole numbers, the combined
spread of the set of Pareto optimal solutions. Crowding dis set 408 would contain better converged and diversified solu
tance d, of solution 210i is equal to distance a, plus by. tions comparing to each independent optimization alone.
For illustration and description simplicity, only two design Finally, dominated solutions are taken out of set 408 to form
objective functions are demonstrated in FIG. 2. Those of 50 the final combined set 420.
ordinary skill in the art would know that other numbers of FIG. 5 is a flowchart illustrating an exemplary process 500
objection functions are used in an optimization conducted for of obtaining such set of diversified solutions by combining
real world projects. Pareto optimal solutions resulted from a plurality of indepen
The goal in a MOEA based engineering design optimiza dently conducted MOEA based engineering design optimi
tion is to obtain a complete global Pareto optimal solution set. 55 zations, according to an embodiment of the present invention.
However, the global Pareto optimal solution set is unknown Process 500 is preferably implemented in software.
for a real-world product, hence a set of solutions that encom Process 500 starts by receiving a description of a product
passes “diversified” solutions is desired. (e.g., consumer electronics, car, airplane, etc.) to be opti
One of the means to achieve such goal is demonstrated in mized at step 502. At step 504, a design space of the product
examples shown in FIGS. 3A-3B. In FIG. 3A, two MOEA 60 is defined by selecting a set of design variables (see FIG. 1A
based optimizations (i.e., first and second optimizations for a very simple example). Also an objective function space
302a-b) of a product are conducted independently. Each opti is defined by identifying a plurality of design objective func
mization starts with a randomly created initial generation of tions in a MOEA based optimization. Next, at step 506, a
populations (i.e., design alternatives). The difference plurality of independent MOEA based engineering design
between these two optimizations is the populations in the 65 optimizations is conducted (see example shown in FIG. 4).
initial generation are different. In one embodiment, different Each of the optimizations is configured to have certain dif
seeds of a random or pseudo-random number generator are ferent characteristics or parameters, for example, design
US 7,996,344 B1
11 12
the engineering design optimizations has converged 9. The system of claim 8, further comprises randomly
based on spread and uniformity measurements, creating the initial generation’s design alternatives for said
wherein the global Pareto optimal solutions are stored each of the engineering design optimization.
in a storage device and graphically displayed in a
monitor upon user’s instruction. #: :#: ::: ::: :#: