JPH01177123A - Element allocation system for hierarchical directed graph - Google Patents

Abstract

PURPOSE:To plot a hierarchical directed graph where the elements are arranged in clusters in an easy-to-see way by deciding not only the vertical direction of each element but the horizontal direction with decision of the hierarchical levels. CONSTITUTION:All series are extracted from a hierarchical directed graph received via a series extracting part 1 and a series weight calculation part 2 calculates the weight of all series. A main series setting part 3 sets the series of the maximum weight at the center as a main series. Then a series selecting part 4 selects the one having the largest weight out of those remaining series and a correlation calculation part 5 obtains the correlation between this selected series and the main series. An addition processing part 6 adds the selected series to the main series. Then a deciding part 7 decides the presence or absence of the series with which the addition processes are not through yet. It the presence of said series is confirmed, the operation sets back at the part 4. While if the absence of such series is confirmed, the relevant procedure is through. Thus it is possible to obtain a hierarchical directed graph where the elements are formed into clusters set in an easy-to-see arrangement.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an element arrangement method for drawing a hierarchical directed graph in which elements are clustered and arranged for easy viewing.

[Conventional technology]

When drawing a hierarchical directed graph, the vertical position of each element [K is determined by the leveling in the hierarchy, but there are no restrictions on the horizontal position, so usually the element number The method used is to arrange them in descending order of age.

[Problem that the invention seeks to solve]

However, this method has the disadvantage that the arrows connecting elements intersect with each other in a complicated manner, resulting in a diagram that is difficult to read.

In order to meet these demands, the purpose of the present invention is to provide an element arrangement method for appropriately arranging elements to obtain a hierarchical directed graph that is easy to see as a whole. It is about providing.

[Failure to solve the problem]

The present invention provides an element placement method for a hierarchical directed graph drawn from n elements and directed lines between the elements defined by an n x n adjacency matrix, from one of the lower end elements to the top element. At the same time as extracting all the paths leading to one sequence (hereinafter referred to as sequences), weights are given to the elements within the sequence according to the level of the hierarchy, and the given weights are summed within each sequence. First, the series with the largest weight is placed in the center as the "main series," and then the next series with the next largest weight is selected. The correlation from the left and the correlation from the right between the next sequence and the main sequence at that point in time are calculated, and a placement process is performed to add from the side with the larger correlation to create a new main sequence.

This is an element arrangement method in a hierarchical directed graph in which a hierarchical directed graph is obtained by repeating the above arrangement process and sequentially adding sequences starting from a sequence with a large weight.

[Effect]

The present invention will be described below with reference to the drawings.

Between elements defined by n elements and n×n adjacency matrix? Regarding graphs drawn using directed lines, the element arrangement method according to the present invention will be described using the hierarchical directed graph shown in FIG. 2 as an example. 1. From the hierarchical directed graph shown in FIG. 2, all paths (hereinafter referred to as sequences) from one of the lower end elements to one of the uppermost elements are extracted. Extracted series? 1 temporary z Pin P2* P3
I will write it as... Next, the weight of the series P is calculated using the following equation.

However, ・eB=l: When the sequence P1 includes an element of the J-th level = 0: Other times al: Total number of levels With this ζ, the values of all sequences and their weights in the example of FIG. 2 are shown in FIG. Shown below. Note that the levels of the elements in FIG. 2 are assigned in order from the top to the bottom of the diagram. Next, the sequence with the largest weight is placed in the center and is called the "main sequence." In the present example, P3 and P6 correspond to this, so they are combined and placed in the center as shown in FIG. Next, select the sequence with the largest weight among the remaining sequences, and calculate the correlation from the left and the correlation from the right between that sequence and the main sequence at that time using the formula (
2) and add them starting from the side with the highest correlation, making this the new main sequence.

Correlation from the left with the main series of series Pi, where fL,
is calculated using the following formula.

rLj ” '8',,%-j,R(8P+'t)R(
8me 81)+・・・・・・・・・ Equation (3) * 3 p: @j level element of series pl・Sm2 Zj level of main series,! ! ! The leftmost element in the set: Su, : Set of elements at a level higher than the mj level @N
up ', Jli prime of set Sup' SLow
: m
w: Number of elements of i combination S LOW - δ + j = l: When i = j = 0: When i door j - R (i, j): Target floor /, reachability matrix of J-like identity graph The (i, j)th component of. Pupil is 1 - t is Oo Also, for the A0 function from the right between series pt and main series, change Sm1 in equation 1 (3) as follows, and change equations (2) and 9 (3). It is sufficient to simply set rLj to rnj. That is.

・Sm: The most right iC element among the @J level elements of the main series In the current example, %P1 corresponds to this, so P2 and the fourth
When the correlation from the left and the correlation from the right with the main sequence in Figure 1 are calculated using equation (2), the correlation from the left is larger, so P2 is added from the left side of the main sequence. The result is as shown in FIG.

Thereafter, the above procedure is repeated to add sequences to the main sequence in descending order of weight, and the process ends when all sequences have been added. FIG. 6 shows the finally obtained hierarchical directed graph.

〔Example〕

FIG. 1 shows a block diagram of an embodiment of the present invention. First, the sequence extraction unit 1 extracts all sequences from a given hierarchical directed graph. Next, the sequence weight calculation unit 2 calculates the weights of all sequences. Next, the main sequence setting unit 3 performs a process of setting the sequence with the maximum weight as the main sequence in the center. Next, the sequence selection unit 4 selects the sequence with the largest weight among the remaining sequences. Next, the correlation calculation unit 5 calculates the correlation between the selected sequence and the main sequence.

Next, the addition processing section 6 adds the series to the main series. Next, the determination unit 7 determines whether there is a sequence that has not been added, and if so, the process returns to the sequence selection unit 4. If not, the process is finished.

Note that it is easy to implement the functions corresponding to the configurations shown in the embodiments using a normal electronic computer.

〔Effect of the invention〕

In contrast to the conventional method of arranging elements in order from the youngest fm number to
This has the effect that a 41i M-directed graph can be obtained.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG.
FIG. 3, FIG. 4, FIG. 5, and FIG. 6 are diagrams for explaining the element arrangement method according to the present invention. 1... Sequence extraction section, 2... Sequence weight calculation section, 3... Main sequence setting section, 4...
...Sequence selection section, 5...Correlation calculation section, 6...
. . . Addition processing section, 7 . . . Judgment section. Agent Patent Attorney Susumu Uchihara Figure 3

Claims (1)

[Claims] In the element arrangement method of a hierarchical directed graph drawn from n elements and directed lines between elements defined by an n × n adjacency matrix, from one of the lower end elements to one of the top elements In addition to extracting all routes (hereinafter referred to as sequences), weights are given to elements within the sequence according to the level of the hierarchy, and the given weights are summed within each sequence to calculate the “weight” of the sequence. First, define the sequence with the largest weight as the “main sequence”.
Next, select the next series with the next highest weight, calculate the correlation from the left and the correlation from the right between the next series and the main series at that time, and select the one with the highest correlation. A method for arranging elements in a hierarchical directed graph, characterized in that a placement process is performed to add from the side to obtain a new main series, and the placement process is repeated to obtain a hierarchical digraph by adding sequentially from the sequence with the largest weight.