 |
Claims  |
|
|
I claim:
1. A method for displaying information in a tree structure on a system
having an input means, a display means and a control means comprising the
steps of:
(a) providing from said input means to said control means input data
representative of information associated with a root of the tree structure
forming a 0.sup.th node level therein;
(b) providing from said input means to said control means input data
representative of nodes on a K.sup.th level in the tree structure where K
is an integer;
(c) receiving input data from said input means representative of a
relationship between said nodes on said K.sup.th level and nodes related
thereto on a different level K+L where L is a nonzero integer;
(d) repeating steps (b) and (c) until information concerning all nodes for
display on the K.sup.th level has been provided to said control means;
(e) operating said control means to generate control data for controlling
said display means to display the nodes of said K.sup.th level and to
display connections representing branches of said tree structure between
said nodes and the nodes related thereto on said different level whereby
any displayed node is displayed together with its related nodes;
(f) repeating steps (b) through (e) as necessary for successive values of K
to represent a tree structure on said display means; and
(g) displaying said nodes of said tree levels and said branch connections
between said nodes and said related nodes,
whereby a tree structure format is provided for displaying said input data.
2. A method as recited in claim 1 wherein said receiving step comprises a
step of choosing L to be a positive integer for receiving data
representative of descendent nodes of said nodes on said Kth level,
said operating step comprises the step of controlling said display means to
display connections between said nodes and said descendent nodes on a
subsequent level, and
said displaying step includes the step of displaying said nodes and said
branch connections between said nodes and said descendent nodes.
3. A method as recited in claim 1 wherein said receiving step comprises a
step of choosing L to be a negative integer for receiving data
representative of antecedent nodes of said nodes on said Kth level,
said operating step comprises the step of controlling said display means to
display connections between said nodes and said antecedent nodes on an
earlier level, and
said displaying step includes the step of displaying said nodes and said
branch connections between said nodes and said antecedent nodes.
4. A method as recited in claim 3 wherein step (b) comprises the step of
providing literal information associated with said nodes on the K.sup.th
level.
5. A method as recited in claim 4 wherein step (g) comprises the step of
displaying said literal information along with said nodes on said display
means.
6. A method as recited in claim 3 wherein step (b) comprises the step of
providing literal information associated with said nodes; and comprising
the further steps of:
providing input data from said input means to said control means indicative
of a particular node on said K.sup.th level to be displayed;
displaying, on a level K+M where M is a nonzero integer, descendent nodes
of said particular node on said K.sup.th level together with literal
information associated therewith;
displaying antecedent nodes of said particular node on the K.sup.th level
in a skeletal format independent of any literal information; and
displaying connections between said particular node and said peer nodes,
said descendent nodes, and said antecedent nodes thereof.
7. A method as recited in claim 6 comprising the further step of selecting
M to equal 1.
8. A method as recited in claim 7 wherein said particular mode and said
descendent nodes comprise the further step of displaying said literal
information associated therewith immediately adjacent said particular node
and said descendent nodes.
9. A method as recited in claim 8 wherein said step of displaying said
antecedent and peer nodes comprises the further step of displaying literal
information associated therewith in a display area remote from said
antecedent and peer nodes.
10. A method as recited in claim 6 wherein said step of providing input
data indicative of a particular node includes the step of providing an
input signal for redrawing an existing prior display of a previously
defined particular node, its peers, antecedents and descendents, and said
steps of displaying antecedent, peer and descendent nodes comprise the
step of replacing said prior display of said previously defined
particularly node, its peers, antecedents and descendents and the
connections therebetween with a display of a newly defined particular
node, its antecedents, peers and descendents and connections therebetween.
11. A method for displaying a decision making algorithm on a computer
activated display having an input means, a processing means and a display
means, comprising the steps of:
(a) providing input data from said input means to said processing means
representative of a target goal to be attained;
(b) providing input data from said input means to said processing means
representative of factors significant in attaining the target;
(c) providing input data from said input means to said processing means
representative of pairwise comparison of significance of each of the
factors in attaining the target;
(d) assigning to each of said factors a numerical value based on pairwise
comparisons of the significance thereof in accordance with a computation
performed on the results of said pairwise comparisons;
(e) providing input data from said input means to said processing means
representative of subfactors significant in each of said factors;
(f) for each factor, providing input data from said input means to said
processing means representative of results of pairwise comparisons of
significance of each of the subfactors for said factor;
(g) assigning to each of said subfactors a numerical value representative
of a local significance thereof to an antecedent factor thereof to which
it is significant in accordance with a computation performed on the
results of said pairwise comparisons;
(h) repeating steps (e), (f) and (g) for sub-subfactors of said subfactors
until all factors and subfactors of interest are accounted for;
(i) displaying on said display means a tree structure including a root, a
plurality of nodes and branches connecting said nodes representing said
target, factors, subfactors, subsubfactors and the like, and descendency
relations between nodes on different levels, respectively.
12. A method as recited in claim 11 wherein said secondary step of
assigning comprises the further step of assigning for display with said
subfactors or the like represented in said display by a node a numerical
value representative of a global significance thereof for attaining the
target goal in accordance with a computation based on the local
significance thereof and a global significance value of a factor
reprsented by an antecedent node thereof.
13. A method as recited in claim 11 wherein said displaying step comprises
the step of displaying at least one of said nodes on said display means of
a K.sup.th level of a tree, along with antecedent nodes on levels K-L is a
nonzero integer, and descendent nodes thereof on levels K+M, where M is a
nonzero integer, together with literal information descriptive thereof.
14. A method as recited in claim 13 further comprising the step of
displaying at least some nodes on said tree structure in skeletal form
including only nodes, branches, and antecedent nodes thereof,
independently of literal information descriptive thereof.
15. A method as recited in claim 11 wherein said first assigning step
comprises the further step of displaying verbal qualitative descriptions
of pairwise relationships for selection and assignment of an appropriate
one of said descriptions to the relationships between each pair of factors
and between each pair of subfactors.
16. A method as recited in claim 11 comprising the further step of
providing different numbers of subfactors for two of said factors.
17. In a method for executing on a programmed computer a process for
selecting among a plurality of alternatives to achieve a desired goal,
including performing pairwise comparisons among a plurality of criteria to
obtain weighting factors for each of said criteria and, for each
criterion, performing pairwise comparisons among said plural alternatives
to obtain numerical values associated with each of said alternatives, and,
based on said comparisons for each criterion, selecting an alternative
having the highest numerical value, the improvement comprising:
a method for generating an output report including the steps of
(a) generating a node grouping having a plurality of nodes and branches
arranged in a tree structure including a root node, corresponding to the
desired goal, connected by a branching arrangement to at least one
descendent node thereof corresponding to one of said criteria, the tree
structure including a plurality of peer, antecedent and descendent
relationships defined by the branching arrangement among the various nodes
thereof,
(b) selecting a particular node as a central node for the report,
(c) providing complete data in association with said selected central node,
(d) providing complete data in association with descendent nodes of said
selected central node,
(e) providing skeletal data, less than the complete data of said selected
central node and said descendent nodes thereof in association with other
nodes of the tree structure, and
(f) generating said output report to display a subset of said nodes and
said branching arrangement of said tree structure, and to display complete
data in association with said central node and said descendent nodes
thereof and skeletal data in association with the other nodes of said tree
structure,
whereby the report is generated with a tree structure in a skeletal form.
18. A method as recited in claim 17 wherein said first and second providing
steps comprise the further step of providing identifying data, branching
data and relative value data associated with said central node and the
descendent nodes of said central node.
19. A method as recited in claim 17 wherein said third providing step
comprises the steps of:
selecting, as said other nodes of the tree structure, peer nodes of said
central node, antecedent nodes of said central node, and peer nodes of
said selected antecedent nodes.
20. A method as recited in claim 19 wherein said third providing step
comprises the step of providing only said branching data in association
with said other nodes of the tree structure.
21. A method as recited in claim 17 wherein said second generating step
comprises the step of displaying a transient output display of said
skeletal tree structure on an output display means of said programmed
computer.
22. A method as recited in claim 17 wherein said second generating step
comprises the step of outputting a report, including said skeletal tree
structure, to an output record medium.
23. A method as recited in claim 22 wherein said outputting step comprises
the step of printing said report on a print carrying medium.
24. Apparatus for displaying information in a tree structure comprising an
input means, a display means, and a control means connected to said input
means and to said display means, said control means being programmed:
(a) for providing input data from said input means to said control means
representative of information associated with a root of the tree structure
forming a 0.sup.th node level therein;
(b) for providing input data from said input means to said control means
representative of nodes on a K.sup.th level in the tree structure;
(c) for receiving input data from said input means representative of a
relationship between said nodes on said K.sup.th level and related nodes
thereto on different levels K+L where L is a nonzero integer;
(d) for repeating steps (b) and (c) until information concerning all nodes
for display on the K.sup.th level has been provided to said control means;
(e) for operating said control means to generate control data for
controlling said display means to display the nodes of said K.sup.th level
and to display connections representing branches of said tree structure
between said nodes and the related nodes thereof on said different levels;
and
(f) for displaying said nodes of said K.sup.th level and connections
between said nodes and said related nodes.
25. A method as recited in claim 24 wherein said control means is further
programmed for choosing said integer L to be a negative integer,
for receiving input data representative of a relationship between said
nodes on said Kth level and antecedent nodes thereof, and
for displaying said nodes of said Kth level and connections between said
nodes and said antecedent nodes thereof.
26. A method as recited in claim 24 wherein said control means is further
programmed for choosing said integer L to be a positive integer,
for receiving input data representative of a relationship between said
nodes on said Kth level and descendent nodes thereof, and
for displaying said nodes of said Kth level and connections between said
nodes and said descendent nodes thereof. |
|
 |
|
 |
Claims |
|
|
|
Description |
|
|
TECHNICAL FIELD
This invention relates to display generation, and more specifically to
display in a hierarchical, tree-like structure, of text and data developed
by a programmed computer. The generated displays include both full and
partial, or skeletal, portions of the tree structure. The invention is
applicable to generation of summarizing reports for a user, whether as a
transient display, on a VDT (video display terminal) for example, or in a
hard copy, printed format.
BACKGROUND ART
Many interactive computer programs are presently available. In such
programs, wherein a user provides input to a programmed processor,
displays are generated by the processor to inform the user of entry of the
inputs. Additionally, the results of computer operation on the input data
are typically displayed for the benefit of the user.
Typically, the output displays are merely in the form of listings of input
or result, whether in the form of numerical or literal information. In
many instances, however, the arrangement of the information to be
displayed is of significance to the user because of the relational
information conveyed by such an arrangement. Frequently such relational
information may be best conveyed by a hierarchical arrangement of the
data, such as found in well known tree-like structure. Illustrations of
the use of such tree structures abound, and include published descriptions
of methods useful in reaching decisions based on a number of criteria, and
also descriptions of organizational hierarchies, as but two examples.
With respect to methods for aiding in decision making, for example, it is
known to select a target goal, a number of criteria to be considered in
choosing among a plurality of alternatives which may be taken to achieve
the desired goal, and possibly to obtain a number of subcriteria which may
be considered in reaching the decision. Weights are assigned to the
various criteria, and an assessment is made with respect to determining
how well each of the alternatives relate to the criteria. Where
subcriteria are involved, the process becomes more complicated in that the
actual weight to be given any of the subcriteria is dependent on the
weight assigned to the criterion with which the subcriterion is
associated.
Finally, each of the alternatives is assessed with respect to each of the
criteria or subcriteria which have been identified, the assessment of the
alternatives being weighted by the weights previously assigned to the
decision making criteria. The alternative having the largest sum of
weighted assessments may be considered to be that alternative most suited
for achieving the desired goal in accordance with the identified criteria.
See, for example, the detailed description of such a process described in
Saaty, Thomas L., Decision Making for Leaders, Wadsworth, Inc., Belmont,
Calif. 1982, incorporated herein by reference. Other algorithms for
assisting in decision making are known.
Clearly, such a decision making process becomes quite complex. In order to
simplify the interaction between the user and the computer, a tree
structure is helpful.
However, none of the interactive systems available in the marketplace,
whether for use in decision making or other applications, provide for
generation of such tree structures. Thus, in any prior art system relating
to a number of complex operations among variables having relational
characteristics, a user is required to generate any relational or
hierarchical chart manually, using paper and pencil in an inexact
operation prone to error.
Additionally, in many applications which benefit from a readily generated
and readily available hierarchical tree chart, it is sometimes feasible to
view only a portion of the chart because of display limitations. Further,
in order to focus a user's attention on particular features and
relationships illustrated by the chart, it is often advantageous to view
only a portion of the chart.
That is, where a tree structure is characterized by a plurality of nodes
interconected by a plurality of branches, it is sometimes desirable to
view only a single subset of related nodes, rather than the entire set of
nodes. Although the partial tree structure may be generated by hand,
machine generation of the same advantageously provides emphasis on a
particular node whenever needed, without the requirement of manual
intervention.
Preferably, such a partial tree structure should include full information
pertinent to a node or nodes of interest, as well as the interconnecting
structure relating the nodes of interest, together with at least minimal
information concerning the remaining nodes of the structure. Particularly,
such a partial tree structure should display the relational
interconnection of at least some of the remaining nodes to the nodes of
interest, but need not display specific data associated therewith.
Such skeletal tree structures are helpful in an organization chart, for
example, in illustrating a chain of command which may exist among a
plurality of departments or offices of the organization. Therein, it maybe
advantageous to indicate the existence of the remaining departments or
offices for the sake of completeness, although complete information is
unnecessary for the omitted departments. It is particularly helpful to be
able to generate such a skeletal chart focusing on any desired department
as its center.
Prior art display generating systems, however, are deficient in the above
described areas and fail to provide either temporary or permanent displays
of tree structures. The prior art fails particularly to provide such tree
structure displays in association with decision making systems implemented
on programmed digital computers.
For example, in a known, commercially available, decision making software
system, there are provided menu driven interactions between the system and
user. The system provides a machine implemented algorithm similar to that
hereinabove described, and includes a number of prompts for the user. Each
user response provides additional data to the system, which accepts and
manipulates the same and performs the necessary computations. However, the
system fails to generate any display to enable the user to assimilate the
various relational information in an integrated manner.
Instead, there is provided only a confirming display of the user generated
input, relating to the various criteria, alternatives, weights to be
assigned to the criteria, etc. Additionally, the results of system
computations are displayed, including ultimate priority to be given to the
various alternatives in accordance with the previously input weights
assigned to the several criteria identified by the user.
It is thus difficult for a user to utilize such a system and to assimilate
quickly and efficiently the relations among the varius criteria,
subcriteria, and alternatives provided to reach the desired goal. It is
particularly difficult to obtain a rapid overview of the
interrelationships among the various variables involved in the problem
being solved.
DISCLOSURE OF INVENTION
It is accordingly a primary object of the present invention to implement a
process for operating a programmed computer to generate a multi-level tree
structure.
It is a more specific object of the invention to provide a machine
generated display of a hierarchical tree structure including a plurality
of nodes and branches connecting said nodes in accordance with
predetermined relationships therebetween.
It is yet another object of the invention to provide a machine generated
display of a tree structure and selectively to provide a skeletal display
of a subset of nodes on branches of the tree, including a complete display
of a first subset of nodes forming a subtree within the tree, together
with all data associated with said first subset of nodes, and a skeletal
display of a further subset of nodes of the tree having a predetermined
relationship to said first subset.
It is an additional object of the invention to provide a skeletal tree
structure display wherein data associated with a first subset of nodes
forming a subtree of a tree are displayed along with said subtree
structure and data associated with a further subset of nodes are omitted,
the skeletal tree display including only the first and further subsets of
nodes and the branches connected thereto.
A further object of the invention is the provision of a machine generated
display of a tree structure as an aid to a user providing input data to a
computer programmed to solve a hierarchically arranged problem.
It is still a further object of the invention to provide a skeletal display
of a tree structure emphasizing a first subset of nodes by providing
descriptive data adjacent a display of said nodes, and to provide
additional display data associated with a further subset of nodes at
locations removed therefrom.
It is another object of the invention to provide, for a computer
implemented decision making algorithm including therein the steps of
selecting verbal or numerical pairwise comparisons among a plurality of
criteria and alternatives of the decision, a machine generated output
display of a tree structure illustrating the relationships among said
criteria and said decision alternatives.
Yet another object of the invention is the provision of a machine generated
skeletal output display for a decision making algorithm.
It is an additional object of the invention to provide a machine generated
hierarchical tree structure display on a transient display and to provide
a permanent copy of such a display in an output report.
In accordance with the foregoing and other objects of the invention, there
is provided a method for displaying information in a tree structure on a
system having an input, a display and a control therefor. The inventive
method comprises the steps of providing input data from the input to the
control representing a root of the tree and the nodes for a K.sup.th level
thereof, together with receiving of data representing a relationship
between the K.sup.th level nodes and antecedent nodes thereof on a level
K-1. The steps are repeated until all the K.sup.th level nodes are
provided to the control, and for several values of K to represent the tree
on the display. The control is operated for causing the system display to
display the nodes of the various levels along with the connections between
these nodes and antecedent nodes thereof, on other levels.
Preferably, the inventive method provides for input of literal information
associated with the various nodes and display of the information adjacent
the nodes. Additionally, the input provides data to the control indicative
of a particular node of the K.sup.th level for display along with its
descendent nodes and literal information associated with each, and
antecedent and peer nodes of the particular node are displayed skeletally,
independently of any literal information associated therewith. Branch
connections between the particular node and the peer, antecedent and
descendent nodes are also displayed.
Moreover, an input signal may be provided for redrawing a tree or skeletal
tree display to emphasize a differently determined particular mode along
with its peers, antecedent and descendent nodes. The information
associated with the particular node and its descendent nodes may be
displayed immediately adjacent these nodes, while the information
associated with the antecedent and descendent nodes may be displayed
remotely from those nodes.
In accordance with another aspect of the invention, there is provided a
method for displaying a decision making algorithm on a computer activated
display having an input, a processor and a display. In the inventive
method, data representative of a target goal are provided to the
processor, along with data representative of factors significant in
attaining the goal. Pairwise comparisons of the significance of the
factors are made and weights are assigned to each of the factors. Similar
comparisons are provided for a number of subfactors associated with the
factors, and weights are assigned thereto. A tree structure is developed
to correspond to the various factors, subfactors, target goal, and various
descendency relationships therebetween, and the tree is displayed on the
display.
In addition to the local significance of the subfactors, the global
significance thereof may be assigned for display in accordance with the
global significance of a factor associated therewith and represented by an
antecedent node thereof.
The display, as previously described, may be complete or skeletal,
including either all or some of the structure of the tree. Moreover,
assignment of the weights to the factors may be based on either verbal or
numerical pairwise comparisons of the significance thereof, based on
qualitative or quantitative criteria. Further, the number of subfactors
associated with any one factor need not be the same as that of any other
factor.
In yet another aspect of the invention, in a process implemented on a
programmed computer for selecting among a plurality of alternatives, a
report is generated by determining a tree structure to include a plurality
of nodes and connecting branches therefor. A central node is selected for
inclusion in the report, and complete data is provided in association
therewith and in association with descendent nodes thereof, while skeletal
data are provided for other nodes of the tree. The other nodes may include
peer nodes of the central node, antecedent nodes thereof, and peer nodes
of the antecedent nodes. The report may be provided on a transient medium
or may be printed on a print carrying medium.
Still other objects and features of the present invention will become
readily apparent to those skilled in the art from the following
description of one specific embodiment thereof, especially when taken in
conjunction with the accompanying drawings.
BRIEF DESCRIPTION OF THE DRAWING
FIG. 1 shows a tree structure including definitions of specific elements
thereof as used in the present invention;
FIG. 2 illustrates a display of a tree structure in accordance with the
invention;
FIGS. 3 and 4 show formats for illustration of a tree structure in a
printed report in accordance with the invention;
FIG. 5 illustrates a skeletal display generated in accordance with the
invention;
FIG. 6 illustrates a sample flow chart broadly descriptive of a method for
implementing the inventive method on a programmed computer;
FIG. 7 shows a detail of FIG. 6;
FIG. 8 illustrates a detailed flow chart for obtaining data for nodes to be
displayed;
FIG. 9 illustrates details of the flow chart of FIG. 6 for generating a
skeletal tree structure; and
FIG. 10 illustrates a detail of the flow chart of FIG. 9; and
FIG. 11 shows an apparatus which may be used to carry out the inventive
method.
Referring now to FIG. 1, a tree structure of the type generated in
accordance with the present invention is illustrated generally at 10. As
is apparent from the figure, the tree is in an inverted form. Included
therein are a plurality of levels, each level having one or more nodes.
At the 0.sup.th level there is provided a root node 12. Descending
therefrom to level 1 are a plurality of descendent nodes. As will be
apparent from the following definitions, the nodes at level 1 form a peer
group of nodes, enclosed for purposes of illustration by a dashed line 14.
Generally, a node on level K will be connected by a branch of the tree
structure to a node on level K-1. Similarly, a node on level K may be
connected by a branch to a subsequent node on level K+M descending
therefrom. Although a tree structure may have branch connections from a
node on level K to a node on level K+M for any value of M, in the
preferred embodiment M is equal to 1. That is, any node N may only be
connected to its direct ancestor or descendent nodes removed therefrom by
one level. Nodes descending from a common node are identified as a peer
group, as illustrated by all the nodes on level 1 and as further
illustrated by a peer group 16 on level 2.
For purposes of clarification, a number of terms utilized in connection
with a tree structure are defined as follows:
TREE
A tree (also called a hierarchy) is a structure which has NODES, and
BRANCHES. The NODES are represented on the display screen and in the
output report by either boxes or small circles, depending on the level of
detail shown for a given part of the tree. Nodes in one level are
connected by branches to nodes in the next levels, either below or above.
Nodes in a tree are organized in levels. The topmost level, Level 0, has
only one node and is called the root of the tree. In a decision making
algorithm, the root represents the GOAL. Each node below Level 0 is
connected to one and only one antecedent node above it (i.e., a node has
only one parent node), but may be connected to many descendent nodes below
(illustratively, a node may have up to seven descendent nodes). Attention
is often focused on a particular node and its descendent nodes.
NODES
Nodes are juncture points in the tree.
LEVELS
All nodes that are descendents of the GOAL node are said to be in Level 1.
Descendents of nodes in Level 1 are said to be in Level 2, etc. Levels are
numbered downward with the GOAL node being at Level 0, nodes in the next
level at Level 1, and so on. Although any number of levels may be
provided, for illustrative purposes the lowest permitted level will be set
at 5.
ANTECEDENT (PARENT)
The node in the next higher level to which a node is connected. A node has
only one parent.
DESCENDENTS (BRANCH, or CHILDREN NODES)
Branch nodes are all the nodes in the next lower level to which a
particular node in a given level is connected. For illustration, each node
may have up to seven descendent or branch nodes.
PEER NODES
These are nodes having the same parent.
LEAVES (ALTERNATIVES)
The nodes at the bottom level of the tree are the leaves. In a decision
making algorithm, the leaves are the alternatives among which the choice
is to be made.
NODE NUMBER
Each node may be identified by a five digit number, one digit for each
possible level below the goal. The K.sup.th digit represents the
left-to-right placement of a node or its antecedent in a group of peer
nodes on the K.sup.th level. Thus, the left-most digit represents the
position (1 to 7) of a node or of its antecedent in Level 1 of the tree.
The second digit represents the position (1 to 7) of a node or of its
antecedent among peers in Level 2 of the tree, and so on. For example, the
node number 22100, shown in FIG. 1, is a descendent of the second node in
Level 1, the second peer in Level 2, and is itself the first branch node
in Level 3. Thus, the first three digits of its node number are 221. The
fourth and fifth digits are 00, as this node is located above the
corresponding Levels 4 and 5.
PAIRWISE COMPARISONS (JUDGMENTS)
These are judgments expressing the intensity of dominance or preference of
one node over another with respect to a given criterion in a decision
making algorithm. Judgments can be expressed either verbally or
numerically.
LOCAL PRIORITIES
Local priorities are the priorities of nodes relative to their peer nodes
as determined by pairwise comparisons.
GLOBAL PRIORITIES
For a decision making algorithm, global priorities are the priorities of
nodes relative to an overall or global decision, determined by weighting
the local priority by the global priority of the parent node.
SYNTHESIZING
Synthesizing is the adding of global priorities of nodes representing a
particular criterion (or alternative), which results in the overall
priority for that criterion (or alternative). Of special interest are the
overall global priorities of the alternatives at the bottom level (the
leaf nodes).
The following illustrates application of the tree structure to a procedure
for deciding on a house to buy. This problem, along with others of varying
complexity, is described in an instruction manual for a decision making
algorithm "EXPERT CHOICE", available from Decision Support Software, Inc.,
of McLean, Va., incorporated herein by reference.
A house-buying decision involves such factors as financial considerations,
geographical location, physical appearance, yard space, neighborhood,
convenience of schools, property taxes, down payment, monthly payment,
commuting ease, and many others. There are usually too many factors in
such a situation to be kept in mind at one time. In accordance with the
invention, a computer implementation of a decision making algorithm
includes a tree structure to enable a user to organize the factors, with
the most important general ones in the top level, detailed breakdowns
thereof in lower levels, and alternatives of choice in the bottom level.
Thus, the entire problem is reduced to smaller, manageable problems. For
demonstration purposes the example shown is small. Many more factors could
be added that are specifically oriented to a particular buyer's needs, and
may be displayed in the tree structure of the invention.
In the house selection problem, the structure might be arranged with the
main goal "Select the Most Desirable House" on Level 0, the major elements
of Financial, Physical and Location on Level 1, more detailed criteria of
these elements on Level 2, and the alternative houses New Development,
Old/Work, and Downtown under consideration on Level 3. The tree structure
of the house-buying example is shown in FIG. 2 and clearly displays the
choice to be made, the main criteria used in making the choice, the
factors significant in each of the main criteria, and the particular
alternatives to be considered.
When presented in such a format the problem is more easily assimilated by
the computer user, and conveys, by the interconnections thereof,
substantially more information than conveyed by a direct listing of the
criteria and subcriteria.
In the structure of FIG. 2, there is provided a single node, identified as
the GOAL, at Level 0 of the tree structure. This node corresponds to the
root node previously defined. At the first level of the tree structure
there are provided the most general criteria of significance to the
problem. For purposes of illustration in the instant problem, these
criteria may be FINANCIAL, PHYSICAL, and LOCATION considerations
associated with the first level nodes.
The nodes on Level 2 are a more detailed breakdown of the elements in the
first level. They are subcriteria used in reaching the goal and together
form the general criteria as follows. Thus, the FINANCIAL criteron
comprises the subcriteria of COST (the initial price), DOWNPAYMENT (the
down payment required), MONTHLY PAYMENT (the monthly cost which covers
everthing else), and TAXES.
The first level node marked PHYSICAL may be composed of SIZE, CONDITION,
and YARD. Finally, the node marked LOCATION is composed of SCHOOLS (their
number and neatness), NEIGHBORHOOD (an assessment of the general quality
of the neighborhood so far as the prospective buyer is concerned), and
TRANSPORTATION (idenifying the quality of transportation services).
On the third level there are illustrated the specific alternatives to be
considered in reaching the desired goal. The houses under consideration
are NEW DEVELOPMENT (representing a tract type house far from town),
OLD/WORK (representing an older home requiring a considerable amount of
restoration), and DOWNTOWN (representing a condominium apartment in the
center of town).
In accordance with the invention the entire tree structure is generated
before entering any judgments, since knowledge of the importance of the
criteria depends on knowledge of the alternatives at the bottom level and
of the grouping and relationships among the various criteria, subcriteria,
and alternatives. A problem model is generated in the form of a tree. The
model also contains judgments, either the initial judgments of equality,
which are automatically entered as the tree is generated, or judgments
entered during previous work sessions. As is apparent from FIG. 2, the
decision making problem is easily comprehended with the aid of a tree
structure, particularly when generated by the system for display to the
user.
With the aid of the present invention, the tree structure of FIG. 2 may be
generated, including therein the various priority values associated with
each of the criteria and subcriteria used in reaching the decision. Such a
structure may be generated interactively on a VDT display and may further
be provided on a hard copy report generated by the system, as shown in
FIG. 3. In an alternative format, the tree structure may be printed in the
form shown in FIG. 4, wherein each node and leaf is shown together with
the global priority thereof and the full textual identification thereof.
Thus, although the tree structure itself may be graphically generated by
the inventive method, as shown in FIG. 3, it is also possible to provide
all the information associated with the tree structure without the graphic
display associated therewith, as illustrated by the report shown in FIG.
4.
In accordance with the invention, there is further provided an output
display, whether in transient form in association with the interactive
transaction or in a printed output format, which highlights a specific
node of interest together with its descendent nodes. Such a display is
shown in FIG. 5, wherein a particular node of interest 20 is identified by
the user. The identification data, which may comprise the node number, are
provided to the display generator. Responsive to receipt of the node
identification and a command to provide a skeletal display, the particular
node is made a central node of the display. That is, the identified node
is displayed as a root node for a subtree structure associated therewith.
Each of the descendent nodes 22 of the central node is displayed,
including textual and numerical information. Additionally, any descendent
nodes of the descendent nodes 22 are similarly displayed. In the
illustration of FIG. 5, obtained from an application of the invention to
yet another decision making problem, the descendents of nodes 22 are the
leaves 24, forming the alternatives for consideration in reaching the
ultimate decision.
Although it is possible to display a complete subtree in which the central
node is in fact made a root therefor, in the preferred embodiment there is
provided a truncated subtree. For reasons of limited display space, it is
only the descendents of the central node and their descendents which are
shown together with full textual information. Thus, in the preferred
embodiment there is provided a skeletal display portion, tracing the
lineage of the central node to the tree root, and a full display portion,
displaying the central node, its "CHILDREN" nodes and its "GRANDCHILDREN"
nodes.
Thus, the present invention provides for generation of a subtree of the
complete tree structure. Additionally, in order to illustrate the proper
placement of the central node within the hierarchical structure, a number
of additional nodes of the tree are displayed. Particularly, in accordance
with the invention all peer nodes of the central node, shown at 26, are
displayed. The antecedent node of the central node and thus of its peers
is similarly displayed as shown at 28. The peer nodes of the antecedent
node are also displayed as shown at 30. Moreover, the antecedent node
thereof is provided at 32, as are its peer nodes thereof. The display thus
shows each of the antecedent nodes of the central node, together with the
peer nodes of any such antecedent node, tracing the descendency of the
central node to the root node 34.
However, unlike the central node and its descendents, the antecedent nodes
and their peers are shown only skeletally, independently of the data
associated therewith. Thus, for the particularly selected central node and
its descendents there are provided full textual data, while for the nodes
related thereto, shown by the various antecedents and peers, it is only
the relational positioning which is displayed.
Such a display is helpful in focusing attention on a specific subset of
nodes, forming a subtree of the tree structure, while maintaining the
significance of the subtree to the entire tree structure in perspective. A
significant advantage of the invention is the ability to display a complex
tree structure in a limited space available in a display. The textual or
numerical description associated with only a s | | |
