 |
Description  |
|
|
The present invention relates to product restructuring, and more
specifically, to a computer implemented process and system for analyzing
product structure to facilitate product restructuring.
RELATED APPLICATIONS
The present application is related to copending commonly assigned U.S.
patent application Ser. No. 07/455,042 entitled "Method and System for
Automated Bill-of-Material Generation" filed Dec. 22, 1989 and commonly
assigned copending U.S. patent application Ser. No. 07/486,704 entitled
"Assembly Table/Selection Table Generation System" filed Feb. 28, 1990.
BACKGROUND OF THE INVENTION
The term "product restructuring" as used herein refers to a process of
generating a listing of assembly components, determining whether the
assembly is being manufactured in a most efficient manner, and if a more
efficient component combination for an assembly appears possible and more
economic, generating a component listing for the apparently more efficient
assembly. The term "efficiency" as used herein may refer to costs in terms
of time, money or other resources. Generally, one assumption utilized in
the product restructuring process is that it is more efficient to use
identical components in many assemblies rather than using different
components in each assembly. More specifically, the cost of manufacturing
an assembly increases as the types and varieties of components increase.
This increased cost leads to inefficient manufacturing processes and
waste.
Limiting the variety of assembly components, however, preferably should not
limit the variety of assemblies manufactured. A motor manufacturer, for
example, prefers to provide a wide range of models. Therefore, product
restructuring should not be performed at the expense of product variety.
In the past, manufacturers have used an assembly matrix to facilitate the
product restructuring process. An assembly matrix comprises a matrix
wherein each column corresponds to an assembly and each row corresponds to
a specific component. An example of an assembly matrix for an electrical
device is shown in FIG. 1.
An assembly matrix generally is utilized to identify deviations in the
selection of types of components for a variety of assemblies. For example,
from a study of the matrix shown in FIG. 1, it is apparent that all 480
volt assemblies use capacitor 2095K92 except the THREE PHASE, 10 Amp
model. The THREE PHASE, 10 Amp model therefore is a good candidate for
product restructuring. Specifically, if capacitor 2095K92 can be used in
this model, capacitor 2095K92 should be used because it will lead to a
more efficient manufacturing process.
Although an assembly matrix conveys much information in compact form, it is
difficult to find all deviations in the matrix, especially when working
with a large matrix involving a large number of assemblies and components,
i.e., a large problem domain. When working in a large problem domain, the
general practice is to decompose the analysis into a number of
sub-matrices. It is time consuming and difficult to generate the assembly
sub-matrices for a large number of assemblies including many components.
Moreover, using a number of sub-matrices hinders a complete product
restructuring analysis. The person performing the restructure process may
never be able to directly compare all assembly components in one matrix
but rather will be limited to sub-matrix comparison. The assembly matrix
method therefore is susceptible to error and is highly dependent upon the
expertise of the person generating and studying the matrices.
Another method of product restructuring is known as "synthesis." The
synthesis method generally requires one or more experts, who should have a
complete understanding of marketing and manufacture, to redesign the
product. During the redesign operation, the experts determine which parts
and products are needed, and the entire array of parts and products are
redesigned from scratch.
Although the synthesis approach may result in massive improvements due to
radical changes, the synthesis approach requires expert knowledge which
may not always be available. Further, the synthesis approach is
susceptible to error due to oversight, and as with the matrix approach, it
is highly dependent upon expert knowledge.
It is therefore an object of the present invention to provide a method and
system for product restructuring which facilitates more economic
generation of products but not at the expense of losing product variety.
Another object of the present invention is to provide a method and system
for product restructuring which is easy to use and implement even for a
large number of assemblies including many components, i.e., in a large
problem domain.
Yet another object of the present invention is to provide a method and
system for product restructuring which is not highly dependent upon the
expertise of a system user.
Still yet another object of the present invention is to provide a method
and system for product restructuring which facilitates a total product
restructure analysis.
SUMMARY OF THE INVENTION
The present method and system for product restructuring provides an
automated manner in which restructure candidates are easily and quickly
identified. A "restructure candidate" is an assembly component identified
as being a component which may possibly be eliminated and/or suitable for
being replaced by a substitute equivalent component.
In accordance with the present invention, assembly tables and
partially-ordered selection criteria tables populated with assembly data
are generated. The selection criteria tables express assembly data as
rules and control selection of components which will result in an assembly
that satisfies predetermined design specifications. More detail with
regard to table generation is provided in copending U.S. patent
application Ser. No. 455,042, titled "Method And System For Automated
Bill-Of-Material Generation", filed Dec. 22, 1989 and incorporated herein,
in its entirety, by reference.
Once the selection criteria tables are populated, the system user may
quickly and easily modify the selection tables by eliminating and/or
substituting selection criteria. During this process, restructure
candidates are identified. Once identified, an expert in the particular
assembly field can determine whether the identified restructure candidates
can actually be replaced or eliminated. The process continues with an
analysis of each restructure candidate and/or until the expert is
satisfied with the product structure.
Accuracy and consistency of the present system are not highly dependent
upon the expertise of a system user. In fact, once selection criteria are
determined, the only point at which a thorough understanding of the
problem domain is necessary is when determining whether an identified
restructure candidate can be replaced or eliminated. Therefore, the system
user can identify the restructure candidates without a need for expert
knowledge. This feature alone may result in a substantial savings in costs
and expert knowledge resources, including time. It should be understood,
however, that expert knowledge is required when determining whether an
identified restructure candidate can actually be restructured.
Importantly, the present invention is easy to use and enables a total
product restructure analysis. Further, the present method and system, by
providing that products can be restructured in an automated manner, also
reduce the time required to bring a restructured product to market. The
present system facilitates more economic generation of products without
sacrificing product variety and provides easy identification of
restructure candidates.
BRIEF DESCRIPTION OF THE DRAWINGS
These and other objects of the present invention, together with further
features and advantages thereof, will be apparent from the following
detailed specification when read together with the accompanying drawings,
in which:
FIG. 1 is an example of a prior art assembly matrix;
FIGS. 2a and 2b are a flowchart illustrating the present system;
FIG. 3 is an example of a hierarchal representation for an electrical
device;
FIG. 4 is an object oriented representation of the electrical device
represented in FIG. 3;
FIG. 5 illustrates a populated assembly table in accordance with the
present invention;
FIGS. 6a-6c illustrate populated sub-assembly tables;
FIGS. 7A-C illustrate selection criteria tables in an electrical device
context;
FIGS. 8A-B illustrate, respectively, the resulting selection tables when
Phase and Amps, respectively, are eliminated as selection criteria for
capacitor selection; and
FIGS. 9A-B illustrate, respectively, the resulting selection tables when
Volts is eliminated as a selection criteria and when Volts is the only
selection criteria for capacitor selection.
DETAILED DESCRIPTION OF THE DRAWINGS
The present invention is directed to a system architecture and method for
automated product restructuring, and the present invention is not directed
to any particular set of data or rules for carrying out a particular
product restructure process. Therefore, although the present invention may
be described herein with reference to, for example, an electrical device,
such example is for illustrative purposes only. The present invention may
be utilized for product restructuring of many products and processes,
including even computer software. Further, the present system architecture
and method are to be practiced on a computer. The present invention,
however, is not limited to practice on one specific computer.
Referring now more particularly to the drawings, FIG. 2 illustrates, in a
flow diagram 200, the sequence of process steps required to make and use
the present product restructuring system. To further facilitate a better
understanding of the present invention, the sequence of process steps is
described herein with reference to an example illustrated in FIGS. 3 to 9.
The example shown in FIGS. 3 to 9 illustrates how the present invention
would be made and used in an electrical device context. The particular
electrical device includes a capacitor, a coil and a resistor. The example
is shown for illustrative purposes only.
Referring now more specifically to FIG. 2, once the process is initiated as
indicated at "START" block 202, an assembly table generation process 204
and a selection table generation process 206 are begun. Although shown as
being performed in parallel, these processes could be performed in series
in any sequential order.
In assembly table generation process 204, and as indicated at block 208, a
system user first models product structure and determines which parts,
i.e. subassemblies, compose a complete assembly. In the electrical device
context, and as shown in FIG. 3, many parts would compose the assembly and
only one level of subassemblies is shown in FIG. 3. Many other
subassemblies are contemplated and could be included. Note that the model
created by the system user may be in one of many forms and, for example,
may even be in a hierarchical tree form described in Ferriter et al., U.S.
Pat. No. 4,847,761.
The next step in the assembly table generation process, as indicated at
block 210, is to convert the assembly model into tables. The preferred
table format results from utilizing a technique known as the Object
Modeling Technique described in Blaha et al., "Relational Database Design
Using An Object-Oriented Methodology," Communications of the ACM, 31, Apr.
4, 1988, which is incorporated herein, in its entirety, by reference.
With the electrical device, the tables which would be created by using the
Object Modeling Technique for the model shown in FIG. 4 are illustrated in
FIG. 5. Each box in FIG. 4 denotes an object class and each object class
corresponds to part of a relational database management system table. The
model includes connectivity in a hierarchal format, illustrated as lines
with arrows at their ends. Specifically, a capacitor, a coil and a
resistor compose the electrical device. A lower portion of the box
labelled ELECTRICAL DEVICE sets forth attributes for that object class.
The specific configuration for the assembly tables for each assembly may be
in one of many forms, and the present invention is not limited to any
particular format. For example, as shown in FIG. 5 and in the electrical
device context, the table labelled "Electrical Device Table" may include
information from the "Electrical Device" class in FIG. 4, and information
related to each subassembly. Each subassembly, as shown in FIG. 6, may
also be represented in a respective table, i.e. a "CAPACITOR TABLE", "COIL
TABLE" and a "RESISTOR TABLE".
The next, and generally final, step in the assembly table generation
process is to load assembly data into the generated tables as indicated at
block 212. If the assembly data is stored electronically, the assembly
table could be populated through an electronic transfer of data. If,
however, the assembly data is not stored in a compatible or electronic
format, the assembly data may have to be manually entered such as through
a keyboard interface or by scanning a printout of the data with an optical
character recognition system.
For the electrical device example, the populated assembly tables which
would be generated are illustrated in FIGS. 5 and 6. In this example,
twenty different electrical devices having different model numbers are
shown. Some of the subassemblies, e.g. capacitor, coil, would include
components which may be utilized in more than one device, and therefore,
less than twenty subcomponents of capacitors and coils would be shown in
their respective tables. For example, and as shown in FIG. 5, capacitor
"1" is utilized in electrical devices 1, 2, 7 and 8. In the capacitor
subassembly table shown in FIG. 6, however, only one entry for capacitor
"1" is required.
Selection table generation process 206, as indicated at block 214, begins
by determining selection criteria. Selection criteria are the factors
which determine a part to be selected, and the selection criteria
generally are elicited from experts in a particular technological field.
In the electrical device context and as shown in FIG. 7A, a capacitor
would be selected based upon the "Phase", "Amps" and "Volts". Therefore,
"Phase", "Amps" and "Volts" are the capacitor selection criteria.
Likewise, and as shown in FIG. 7B, selection criteria for a coil is
"Phase", and as shown in FIG. 7C selection criteria for a resistor are
"Phase" and "Amps". Selection criteria are determined for each
subassembly.
Once the selection criteria are determined, the criteria are converted into
a table format as indicated at block 216. At this point, the selection
tables would be empty, i.e. no assembly data in the tables. In the
electrical device context, the selection criteria model which would be
created is shown in FIGS. 7A-C. It should be apparent that while the
assembly model tables would describe an electrical device, the selection
criteria tables provide a format for choosing an electrical device.
The next step, as indicated at block 218, is to determine a partial
ordering for the selection tables. This partial ordering must be
determined on an application specific basis. For example, in a motor
context, one subassembly of a motor is an end shield. Generally, screws
are utilized to mount the end shield to another part of the assembly.
Before the type and number of screws can be selected, the type of end
shield must be known. Therefore, the end shield selection process must be
performed prior to the screw selection process. The partial ordering
process may be performed by a human operator or may be automated. An
automated partial ordering process, for example, may be performed by a
computer implemented process. Expert knowledge generally is not required
to perform the ordering process. This partial ordering process is not
necessary for practice of the present invention.
As a result of the above-described selection table generation process, a
sequence of empty, partially ordered, selection database tables are
generated. Further, as a result of the assembly table generation process,
populated assembly database tables will have been generated. The next
step, as indicated at block 220, is to compute selection rules, i.e.,
generate populated selection criteria tables. This step may be performed
by determining all possible combinations of subassemblies utilizing the
assembly table and selection table. For example, referring to FIG. 7A and
in an electrical device context, the capacitor selection table actually
would be an expression of a logic rule. In the capacitor selection
operation, the rule would be expressed as:
If "Phase"=A and "Amps"=B and "Volts"=C,
then "Capacitor Number"=D.
For example, if the Phase is "3", Amps is "10" and Volts is "220", then the
selected capacitor will be capacitor number "2". Values for Phase, Amps
and Volts are variable, and may be referred to herein as selection
criteria variables. By utilizing the capacitor rule, and scanning the
assembly table for all encountered combinations of the rule elements, e.g.
Phase, Amps and Volts, the criteria selection tables would be populated.
More detail with regard to the table generation processes is provided in
copending U.S. patent application Ser. No. 455,042, titled "Method And
System For Automated Bill-Of-Material Generation", filed Dec. 22, 1989 and
as pointed out above, incorporated herein, in its entirety, by reference.
The process, up to this point and as described above, may require a
relatively long time period, e.g. weeks, to complete as compared to the
time required for the following described product restructure processes,
e.g. minutes. More specifically, the time required to generate populated
assembly and selection criteria tables is greater than the time required
to perform the restructure analysis of these tables. The time required to
generated populated assembly and selection criteria tables depends upon
the specific problem domain. Importantly, although the table generation
process may require a relatively long time period to complete, the process
need only be performed once. Therefore, the overall restructure process
may require less time to complete than other known restructure processes.
The next step, as indicated at block 222, is to try deleting a selection
criterion from a selection table. For example, in the electrical device
context, the coil and resistor selection tables, FIGS. 7B and 7C, appear
compact, so there does not appear to be much potential restructuring of
these components. The capacitor selection table, FIG. 7A, on the other
hand, appears relatively long. Therefore, the capacitor selection table
most likely will include some restructure candidates.
As shown in FIG. 8A, and as a first hypothesis, the system user may try
deleting Phase as a selection criterion from the capacitor selection
table. The selection table is then scanned to determine a modified
selection table. Specifically, the remaining selection criteria, e.g. Amps
and Volts, are combined where identical to correspond to one row. For
example, FIG. 8A illustrates the results if Phase is not used as a
selection criteria. The "2.sup.2 " in the row corresponding to "10 Amps"
and "220 Volts" means that capacitor number 2 is selected at both
occurrences of "10 Amps", "220 Volts".
Based upon the results shown in FIG. 8A, it should be apparent that Phase
is not a critical selection criteria in capacitor selection. In fact, the
only role of phase is to resolve the case of 10 Amps, 480 Volts. Capacitor
3 is used in a first 10 Amp, 480 Volt assembly and capacitor 4 is used in
a second 10 Amp, 480 Volt assembly.
A next step in the restructure analysis, as shown in block 224 in FIG. 2,
is to review parts for possible elimination. Determining whether a
component may be eliminated or whether a substitute part may be utilized
requires an understanding of the problem domain. Identifying these
restructure candidates, however, does not require expert knowledge. For
example, from the results shown in FIG. 8A, capacitor numbers 3 and 4 are
identified as restructure candidates, i.e., it may be possible to use
either capacitor number 3 or capacitor number 4, rather than capacitor
number 3 for one 10 Amp 480 Volt configuration and capacitor number 4 for
another 10 Amp, 480 Volt configuration. Determining whether a substitution
may be made, however, requires an understanding of the problem domain.
After the identified restructure candidates are identified and analyzed,
and as shown in block 226 in FIG. 2, the system user/expert determines
whether the present product structure is satisfactory. If the system
user/expert is satisfied with the structure, then operations are stopped
as indicated by an end block 228.
If, however, the system user is not satisfied with the product structure,
the system user may continue operations by again modifying the selection
criteria table by deleting a different criteria from the table as shown in
block 222. For example, FIG. 8B illustrates the results of this step in
the electrical device context. In FIG. 8B, Amps has been deleted as a
selection criterion from the capacitor selection table From FIG. 8B, it is
apparent that the only role of the "Amps" criteria is to resolve the case
of the "3 Phase" "480 Volt" configuration. It also is apparent that
capacitors number 3 and 4 may be suitable for restructuring, as was also
evident from the selection criteria table when "Phase" was taken out as a
selection criteria.
As shown in FIG. 9A, the system user then tried deleting "Volts" as a
selection criteria in capacitor selection. From the results shown in FIG.
9B, it is apparent that "Volts" is an important selection criteria in
capacitor selection because it is needed to resolve which "Capacitor
Number" to select for every configuration.
From the foregoing process of deleting selection criteria and identifying
restructure candidates, the system user then reasoned that "Volts" may be
the only necessary selection criteria and these results are shown in FIG.
9B. As is apparent from FIG. 9B, the only row where more information is
needed to determine the "Capacitor Number" is the row corresponding to
"480 Volts", and this row includes the previously identified restructure
candidates. This selection table also evidences that Capacitor Number 3 is
more popular than Capacitor Number 4. That is, Capacitor Number 3 is
utilized in three configurations whereas Capacitor Number 4 is utilized in
only one configuration. In accordance with the previous explained
assumptions of product restructing, Capacitor Number 4 is the best
candidate for product restructuring. Once identified, a system user with
an understanding of the electrical device may decide whether Capacitor
Number 4 can be eliminated and, for example, Capacitor Number 3 used in
its place.
The processes of the present invention may, for example, be performed on
and/or by a computer known in the art as a VAX 11/785, and generally are
controlled by a human system user. The present invention may be utilized
with most all commercially available computer systems, and it is
contemplated that artificial intelligence techniques may be utilized in
combination with the human-performed steps.
From the foregoing discussion, it should be apparent that the present
system and method provide easy identification of restructure candidates.
The selection table analysis may be extended and effectively utilized in a
large problem domain and is easy to automate. It should be understood,
however, that the system is an adjunct and stimulus to human decisions in
the manufacturing process. Specifically, a person with an understanding of
the problem domain must determine whether identified restructure
candidates may actually be restructured.
The present invention also facilitates a complete restructure analysis. The
step-by-step procedure of the present invention is easy to implement, and
does not require expert knowledge for identifying restructure candidates.
Design accuracy and consistency of the present system are not highly
dependent upon the expertise of a system user. Although an understanding
of the problem domain is required at one point, the present invention
operates more on a "trial and error" basis than being dependent upon
expert knowledge. The present system therefore reduces errors which may
result from inconsistent design techniques. Also, in order to implement a
new design technique, the present system can be easily modified so that
the stored data and rules conform to the new design technique.
The present method and system, by providing that products can be
restructured in an automated manner, also reduces the time required to
bring a restructured product to market. The present system also
facilitates more economic generation of products without sacrificing
product variety.
While an embodiment has been illustrated and described herein, it will be
apparent that numerous modifications, changes, variations, substitutions,
and equivalents, in whole or in part, will now occur to those skilled in
the art without departing from the spirit and scope contemplated by the
invention. Accordingly, | | |
