Ex Parte Casati et al

Board of Patent Appeals and Interferences

Dec 6, 2010

10177423 (B.P.A.I. Dec. 6, 2010)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________

BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES
____________

Ex parte FABIO CASATI MEHMET SAYAL, MARIA GUADALUPE
CASTELLANOS and DIMITRIOS GUNOPULOS
____________

Appeal 2009-010990
Application 10/177,423
Technology Center 3600
____________

Before: MURRIEL E. CRAWFORD, ANTON W. FETTING, and BIBHU
R. MOHANTY, Administrative Patent Judges.

CRAWFORD, Administrative Patent Judge.

DECISION ON APPEAL
Appeal 2009-010990
Application 10/177,423

2
STATEMENT OF THE CASE
This is an appeal from the final rejection of claims 1-26. We have
jurisdiction to review the case under 35 U.S.C. §§ 134 and 6 (2002).
The claimed invention is directed to systems and methods for
analyzing decision points in business processes (Spec. 1:12-13). Claim 1,
reproduced below, is further illustrative of the claimed subject matter.
1. A method, comprising:
accessing process execution data
characterizing execution of one or more instances
of a business process involving a set of one or
more activities and a set of one or more decision
points; and
based upon the accessed process execution
data, building a predictive quantitative model
comprising a set of one or more correlation rules
correlating context data associated with respective
stages of the business process with possible
outcomes of the business process from respective
ones of the decision points in the business process.

Claims 1, 13-15, and 24 stand rejected under 35 U.S.C. § 102(b) as
anticipated by Alexander (US Pat. 5,182,793, iss. Jan. 26, 1993); claims 2-4
stand rejected under 35 U.S.C. § 103(a) as unpatentable over Alexander in
view of Du (US Pat. 6,041,306, iss. Mar. 21, 2000); and claims 5-12, 16-23,
25, and 26 stand rejected under 35 U.S.C. § 103(a) as unpatentable over
Alexander in view of Ullman (US Pat. 6,631,362 B1, iss. Oct. 7, 2003)1.
We AFFIRM.

1 The rejection of claims 15-23 under 35 U.S.C. § 101 was withdrawn by the
Examiner. (See Exam’r’s Ans. 2).
Appeal 2009-010990
Application 10/177,423

3
ISSUES
Did the Examiner err in asserting that claims 1, 13-15, and 24 are
anticipated by Alexander?
Did the Examiner err in asserting that claims 2-4 are unpatentable
over Alexander in view of Du?

FINDINGS OF FACT
Alexander
Alexander discloses that an example of an application of the apparatus
is a computer network for monitoring an airline flight operation, in which
the domain of the operation includes all events affecting flight schedules.
As will be explained below, the system provides real time support for
enabling a user to comprehend the scope of a problem, observe the details of
problem side effects, generate multiple possibilities for improving the
situation, and evaluate alternatives for improvement. The use of this specific
application is exemplary, and the apparatus and method of this invention
could be used for any number of other applications in many diverse areas
(col. 3, ll. 13-24).
Decision processor system 60 includes two subsystems, including an
effects processor 62 and a strategy processor 66, each associated with
special programming to accomplish the functions described below. The
hardware associated with each processor may be any one of a number of
well-known devices capable of executing computer instructions, such as a
microprocessor (col. 4, ll. 33-39).
Effects processor 62 embodies the concept of a “‘symbolic
spreadsheet.’” The symbolic spreadsheet is analogous to numeric
Appeal 2009-010990
Application 10/177,423

4
spreadsheets in common use today, in which numbers or formulas are
assigned to cells. The user can change a value in one cell and immediately
see the effect on values of other cells. In contrast to numeric spreadsheets,
effects processor 62 uses frames as cells to symbolically represent
components of a particular domain. Values, complex entities, or even
programs may be assigned to cells. The frames are linked by means of
descriptions of their relationships. When a change in the domain occurs or
when a hypothetical event is proposed by a user, the programming
determines the effect of those changes on other aspects of the operation. A
term used to refer to this type of programming is “‘constraint propagation’”
(col. 4, ll. 40-55).
For the airline flight operations example, cells represent objects, such
as airplanes, crews, and airports, and events that affect them, such as flight
delays, airport closings, or maintenance delays. When the system receives
an event, effects processor 62 determines its effects and updates knowledge
base 50 (col. 4, ll. 56-61).
A second subsystem of decision processing system 60 is the strategy
processor 66. The programming of strategy processor 66 permits several
basic functions: solution of “‘best choice’” problems using a particular
method, variation of methods, and variation of strategies (col. 5, ll. 3-8).
Figures 3-7 illustrate another aspect of the invention, which is a
method for using a computer to assist a user in selecting a best choice among
alternative choices when making a decision. Essentially, Figures 3, 4A, and
4B illustrate the method of making a best choice, given a particular choice
method. Figures 5 and 6 illustrate the method of the invention using strategy
selection and learning features. Figure 7 illustrates a method of the
Appeal 2009-010990
Application 10/177,423

5
invention using a scenario feature. These methods can be implemented with
the functions and data types discussed in connection with Figure 2 (col. 8, ll.
56-67).
As shown in Figure 5, using the invention with the strategy selection
features involves a first step, Step 510, of adopting a strategy. When a
strategy is adopted, the parameter values and choice methods of that strategy
become the current values and choice methods. Step 512 is executing the
decision making steps of Figure 3 to choose a best candidate. Step 514
determines if the best candidate is acceptable. If it is, Step 516 is using the
best candidate in other programming or as desired by the user (col. 12, ll. 59-
68).
Steps 518 and 520 of Figure 5 illustrate another feature of the
invention, which may be used during the development stage of the decision
making system. This is a learning feature, which is invoked by the user after
a decision making operation has failed to produce the desired result. In
general, the learning feature is a rule weight adjusting feature in connection
with the weighted choose method, so that the candidate who is the user's best
choice will receive the highest score, and thereby improve the knowledge
base. The learning steps are iterative for each “‘rule instance,’” i.e., each
part of the weighted choose method in which a particular rule was evaluated
for the entire set of candidates surviving elimination. A rule instance is the
rule weight and a list of candidate/rank weight pairs (col. 13, ll. 1-15).
Step 518 determines whether this learning feature is to be invoked. If
so, Step 520 is performing it. The substeps of Step 520 are illustrated in
Figure 6. It is assumed that the programming has already saved each
candidate and its state information, including the total score for each
Appeal 2009-010990
Application 10/177,423

6
candidate, and each rule instance. Step 610 is getting a rule instance. Step
612 ensures that the learning steps continue for each rule instance (col. 13,
ll. 16-23).
Step 614 is determining the relationship between the current rule and
the current candidate's total, T. For each rule, there are three status
categories: (1) every candidate that has a higher current T than the desired
candidate is ranked higher by the current rule, or (2) every candidate that has
a higher current T than the desired candidate is ranked lower by the current
rule, or (3) neither category (1) nor (2) exists. According to the category in
which a rule falls, the remaining steps increase or decrease that rule's weight
or set the rule aside (col. 13, ll. 24-34).
If the rule falls within category 3, Step 618 sets the rule aside. Steps
620-626 ensures that if the rule is set aside, after any rule that falls within
category (1) or (2) has had its weight adjusted, the set aside rules can be
looked at again to determine if any one of them now fall within category (1)
or (2) (col. 13, ll. 35-40).

Du
Du discloses that organizations are currently engaging in workflow
process re-engineering in many domains, including financial services,
telecommunications services, healthcare services, customer order
fulfillment, manufacturing procedure automation, and electronic commerce
(col. 1, ll. 34-39).
Figure 3 shows, by way of example, a workflow process 18 which is
represented as a directed graph 40 consisting of a set of nodes connected by
arcs as displayed on the HP OpenPM user interface (col. 6, ll. 17-20).
Appeal 2009-010990
Application 10/177,423

7
Associated with each workflow process 18, there is a process data
template defined by a workflow process designer module 22a (shown in
Figure 2). The process data template is used to provide initial data for the
creation of process instances. At run time, based on the process data
template and read/write lists of activities defined in a workflow process 18,
HP OpenPM will generate a case packet for each process instance to
facilitate data passing between activities and the HP OpenPM engine 20
(col. 7, ll. 25-34).

Ullman
Ullman discloses in Figure 12 that the box labeled “‘Decision’” 52
takes as values the alternatives for resolving the issue represented by the
diagram. The circle labeled S(Cc│A a) 53 represents the satisfaction of
criterion Cc given alternative A a and will be called a satisfaction node.
While we show only one, there can be one for each alternative/criterion
combination. This is nonstandard, as a usual influence diagram would have
only one node for each criterion. We use contingent nodes to represent each
alternative/criterion pair separately, since we may not have arguments for
every pair. The pair of two-node chains 54 hanging from S(Cc│A a)
represents opinions posted by decision-makers. Two are shown here, but
there can be any number of such chains hanging from each of the S(Cc│A a)
satisfaction nodes. The higher of the two circles 55 represents the state of
participant knowledge about the ability of the alternative to meet the
criterion, and the lower circle 56 is a diagram artifact used to encode
probabilistic evidence (col. 5, ll. 20-37).
Appeal 2009-010990
Application 10/177,423

8
ANALYSIS
Claims 1, 13-15, and 24
We are not persuaded that the Examiner erred in asserting that claims
1, 13-15, and 24 are anticipated by Alexander (App. Br. 11-22). By saying
that claims 13-15 and 24 are patentable for the same reason as independent
claim 1, Appellants have de facto chosen independent claim 1 as being
representative of the subject matter of claims 1, 13-15, and 24 (App. Br. 19-
22; Reply Br. 7-8). We do the same.
Appellants assert that Alexander does not disclose a method that
operates with respect to a single business process, as recited in independent
claim 1 (App. Br. 14-15, 18; Reply Br. 2-3). First of all, Alexander
discloses methods applied to flight operations and decision selection (col. 3,
ll. 13-16; col. 4, ll. 56-59; col. 5, ll. 3-8; col. 8, ll. 56-59). These are
business processes under a broadest reasonable interpretation. See In re Am.
Acad. of Sci. Tech. Ctr., 367 F.3d 1359, 1364 (Fed. Cir. 2004). Moreover,
Alexander also discloses that the “specific application is exemplary, and the
apparatus and method of this invention could be used for any number of
other applications in many diverse areas” (col. 3, ll. 21-24).
Furthermore, the fact that different structures in Alexander may
perform different portions of the business process does not mean the
different portions are not a part of the same “single” business process. And
in any case, the requirement of a single business process performed by a
single structure is not set forth in the claim. See CollegeNet, Inc. v.
ApplyYourself, Inc., 418 F.3d 1225, 1231 (Fed. Cir. 2005).
Appellants assert that Alexander does not disclose a predictive
quantitative model, as recited in independent claim 1 (App. Br. 15, 17;
Appeal 2009-010990
Application 10/177,423

9
Reply Br. 3-4). However, the Examiner has cited numerous portions of
Alexander which disclose that inputs are used to make predictions on
changes and effects, for example, with respect to airline operations (col. 4, ll.
50-54, 59-68; col. 5, ll. 3-8). Furthermore, choice selection, as disclosed in
Alexander, is also a prediction of the best scenario (col. 8, ll. 56-59; col. 10,
ll. 27-30; col. 12, ll. 53-56).
Appellants assert that Alexander does not disclose
a set of one or more correlation rules correlating
context data associated with respective stages of
the business process with possible outcomes of the
business process from respective ones of the
decision points in the business process[,]
as recited in independent claim 1 (App. Br. 16, 18; Reply Br. 4-5). As an
initial matter, we note that altering an input resulting in an altered output has
a perfect correlation (i.e., correlation coefficient of one) under a broadest
reasonable interpretation. Along those lines, Alexander discloses that
“[w]hen a change in the domain occurs or when a hypothetical event is
proposed by a user, the programming determines the effect of those changes
on other aspects of the operation” (col. 4, ll. 50-54, 59-61). Thus, the
“hypothetical event” is “predictive,” and the “changes on other aspects of
the operation” in Alexander corresponds to the recited “possible outcomes of
the business process from respective ones of the decision points in the
business process.”
Appellants assert that the weights and links for inputs are pre-
programmed, and thus are not capable of being dynamically changed “based
upon the accesses process execution data,” as recited in independent claim 1
(App. Br. 16, 18; Reply Br. 5-7). However, Alexander discloses that
Appeal 2009-010990
Application 10/177,423

10
“[Figures] 5 and 6 illustrate the [] method of the invention using strategy
selection and learning features” (col. 8, ll. 61-63; col. 12, ll. 59-63; col. 13,
ll. 1-40). Specifically, Alexander discloses a learning feature that adjusts the
weights, which dynamically changes the correlations based on data, as
claimed.

Claims 2-4
We are not persuaded that the Examiner erred in asserting that claims
2-4 are unpatentable over Alexander in view of Du (App. Br. 24-27; Reply
Br. 9-10). By saying that claims 3 and 4 are patentable for the same reason
as claim 2, Appellants have de facto chosen claim 2 as being representative
of the subject matter of claims 2-4 (App. Br. 27; Reply Br. 10). We do the
same.
Appellants assert that Du does not disclose “partitioning the business
process into the stages,” as recited in dependent claim 2 (App. Br. 25).
However, the nodes and process instances disclosed in Du correspond to the
claimed stages (col. 6, ll. 17-20; col. 7, ll. 25-34).
Appellants assert that there is no reason to combine Alexander and Du
(App. Br. 25-27; Reply Br. 9-10). However, Du discloses that
“[o]rganizations are currently engaging in workflow process re-engineering
in many domains,” of which creating nodes and process instances are stages
(col. 1, ll. 35-39). Thus, one of ordinary skill would implement the stages of
Du in Alexander to re-engineer Alexander, as the Examiner puts it, “to
provide a more accurate description of the procedural knowledge and a more
efficient workflow analysis,” as desired by Alexander (col. 1, ll. 33-38).

Appeal 2009-010990
Application 10/177,423

11
Claims 5-12, 16-23, 25, and 26
We are not persuaded that the Examiner erred in asserting that claims
5-12, 16-23, 25 and 26 are unpatentable over Alexander in view of Ullman
(App. Br. 27-33; Reply Br. 10-13). By saying that claims 6-12, 16-23, 25,
and 26 are patentable for the same reason as claim 5, Appellants have de
facto chosen claim 5 as being representative of the subject matter of claims
6-12, 16-23, 25, and 26 (App. Br. 31-33; Reply Br. 12-13). We do the same.
Appellants assert that Ullman does not disclose that
each of the correlation rules specifies respective
probabilities of the possible outcomes of the
business process from a respective one of the
decision points conditioned on completion of all
activities of the business process up to a respective
one of the stages of the business process and a
respective context data value[,]
as recited in claim 5 (App. Br. 29-30; Reply Br. 11-12). However, Ullman
discloses multiple satisfaction nodes at different points in a process
dependent on different alternatives/criterion (col. 5, ll. 20-37). The content
of the satisfaction nodes corresponds to the recited possible outcomes under
a broadest reasonable interpretation. See In re Am. Acad. of Sci. Tech. Ctr.,
367 F.3d at 1364.
The decision of the Examiner to reject claims 1, 13-15, and 24 under
35 U.S.C. § 102(b) is AFFIRMED.
The decision of the Examiner to reject claims 2-4 under 35 U.S.C. §
103(a) is AFFIRMED.
The decision of the Examiner to reject claims 5-12, 16-23, 25, and 26
under 35 U.S.C. § 103(a) is AFFIRMED.
Appeal 2009-010990
Application 10/177,423

12
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a). See 37 C.F.R. §
1.136(a)(1)(iv) (2007).

AFFIRMED

hh

HEWLETT-PACKARD COMPANY
Intellectual Property Administration
3404 E. Harmony Road
Mail Stop 35
FORT COLLINS, CO 80528