JPS6152036A - Switch control system of spatial split type switch - Google Patents

Abstract

PURPOSE:To control the number of times of switching of a spatial split type switch by obtaining switch changeover pattern where the frame length is maximized and the number of times of switching is minimized even when any traffic demmand matrix is given by diversified call information. CONSTITUTION:A traffic at each area is collected by a call information management section 2 based on call information from each station inputted from a call information input terminal 1 and the traffic demmand matrix D between areas is generated at a traffic demmand matrix generating section 3 based on the traffic information. A switch changeover pattern setting section 4 expands the matrix into the set of basix matrixes (a matrix including at most one non-zero element at each row and each column). The pattern given by the set is stored in a switch changeover pattern memory section 5. A switch changeover control section 6 switches a reception data burst (traffic) from a reception section 9 to a path corresponding thereto based on the stored pattern to control a spatial split type switch section 7 and the result is outputted to a transmitter 10.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention provides a switching system including a space-division type switch, in which the amount of traffic applied to a plurality of input terminals of the switch is divided into two or more predetermined outputs of the switch. Related to a switching control method for a space-division switch that minimizes the time (frame length) required for transmitting the amount of traffic and also minimizes the number of switch changes when controlling the switching for guiding and transmitting traffic to a terminal. It is.

[Prior art and its problems]

In a switching system using a space division type switch (hereinafter also simply referred to as a space switch), when controlling the switching of the switch, the form of traffic applied to the switch (
It is important to appropriately determine the switching pattern of the switch according to the amount of data and the destination (destination) in order to minimize the frame length and the number of times the switch is switched.

As an example of switch switching control in a switching system using space switches, SS-TDMA (Satellite
5w1tched -Time Divisio
An example of this is switching control of satellite switches in the nMultiple Access) system.

Figure 6 is a conceptual diagram of the SS-TDMA system. In the figure, 100 represents a communication satellite, Ai represents a spot beam area covered by the same frequency, Xi represents an uplink, and Y
i indicates the downlink (where i=1.2...
・n).

In this 5S-TD, MA system, each beam area (Al
When transmitting a data burst (a mass of time-division multiplexed information) as traffic from the target area (An if An), space division is required to switch and connect this traffic from Xl to Yn on the satellite 100. The switch is Pf&.

In this way, the traffic distribution currently occurring between each beam area is summarized in a matrix as shown in FIG. 7, which is called a traffic demand matrix.

That is, in FIG. 7, for example, the element d(LL) is connected from the uplink X1 in the area A1 to the communication satellite 10.
Similarly, for example, the element d(2,i) indicates the traffic demand from WX2 to the downlink Y1 in the same area AI through the communication satellite 100. shows the traffic demand amount to the downlink Yi.

This means that the input terminals of the space switch provided in the communication satellite 100 are X1, X2. ...Xi...n of Xn
There are also output terminals Y1, Y2. ...Yi...Yn
There are n terminals, which means that switching control must be performed between these input and output terminals to satisfy the traffic demand as shown in FIG.

However, if d(1
, 1), the line Y
Therefore, even if there is a traffic demand of d(2,1) from another uplink line X2 to line Y1, transmissions that satisfy this demand cannot be performed simultaneously. The same thing can be said about other lines.

Therefore, it is conceivable that there are a number of spatial switch switching patterns that indicate from which uplink to which downlink, and when, transmission that satisfies the traffic demand should be performed by switching.

FIG. 8(A) is an explanatory diagram of a switching pattern showing an example thereof.

In the figure, the spot beam area is Al.

A2. In the case of A3, the uplink is XI.

X2. The downlink is the same for all three of X3 <Y1, Y2゜Y3
We assume three cases.

First, let's focus on uplink x1. Uplink x1 occupies downlink Y2 for the period of time slot ■, and then there is an empty period (time slots ■, ■) shown by the shaded area.
Next, connect down line Y3 to ■, ■.

It occupies the downlink Y1 for the time slot period of ■, then there is an empty period (time slot ■) shown with diagonal lines, and then it occupies the downlink Y1 for the time slot period of ■, ■, [phase]. , the entire traffic demand from uplink X1 is satisfied and one frame ends.

Other uplink X2. Similarly, for X3, the downlinks Y1, Y2, . Each must occupy Y3 and satisfy its own traffic demand.

In FIG. 8(A), X1, X2. It will be appreciated that there is no period during which each uplink of X3 occupies the same downlink at the same time.

For uplink XI, switch switching is performed between time slots ■ and ■, between ■ and ■, between ■ and ■, and between ■ and ■.
It occurs at the end of [phase].

Regarding uplink X2, similarly, between time slots ■ and ■, between ■ and ■, between ■ and ■, between ■ and [phase], and [
occurs at the end of phase]. There is no need to explain the upstream vAX 3.

Next, please refer to FIG. 8(B). This pattern includes uplinks X1, X2 . Downlinks Y1, Y2 . Y3
It will be seen that while satisfying exactly the same traffic demand as in (A), the frame length only needs to be the length of tie J slots ■ to ■, and the number of switchovers is significantly reduced. .

This is because the pattern arrangement has been devised to eliminate the empty periods caused by the hatched areas seen in (A).

From the above, given the traffic demand matrix shown in Fig. 7, it is necessary to develop a method for finding the optimal pattern shown in Fig. 8 (B) instead of the pattern shown in Fig. 8 (A). It will be recognized that it is important.

One such existing method is an extraction method for obtaining the optimal approximate solution of switch switching patterns (Iko, Urano, Muratani, "Optimal allocation of time slots in rSS/TDMA system", 'Junko Communication Society paper) Magazine (B) Im
p 98-105, Vol 61-B.

According to this method, the optimal approximate solution pattern is obtained according to the following procedure.

(Step 1) Find the row (column) in which the sum of the amounts indicated by each element in each row or column in the given traffic demand matrix D++1, that is, the row (column) sum, is maximum. And i. The th row (column) is called the 1st allegory row (column).

(Step 2) The element showing the maximum amount (maximum element) is extracted from this critical row (column) as the critical element dc.

(Step 3) Exclude the row and column in which the Hm field element dc is included, and extract the largest element from the remaining elements of the matrix.

(Step 4) Perform s on the remaining matrix excluding the rows and columns containing the elements found in Step 3 and earlier.
Repeat step 3 (n-1) times.

(step 5) n found up to 5step 4
elements (one element selected from each row and each column) are extracted from the traffic demand matrix D″′, and the basic matrix D1 (
1) and the remainder to create the residual matrix D 2
”.In other words, the following equation holds true.

D +11 = p, +11 −
D 2 (11 (step 63 basic matrix D
1 (1) and the 10 rows (columns) of the residual matrix D2'' are critical rows (columns) in each row (column) (row sum or subphase or maximum), otherwise go to steplO. 5 step 7”, line (.

(step ? 'J fundamental matrix D1 (1) critical element d
If there is an element dm that indicates a larger amount than c, the excess of dm is returned to the corresponding element of the residual matrix D%11 until both are equal.

(Step 8) i of residual matrix D2(1). queue)
10 rows (columns) if is no longer a critical row (column)
5te until becomes the critical row (column) of the residual matrix D2(1)
While satisfying p7, return one unit at a time from the basic matrix D+L' to the corresponding elements of the residual matrix D2+11 (guaranteeing the critical condition of the residual matrix D%11).

(Step 9) When the residual matrix DI'' becomes a zero matrix, that is, when D(1) = D%I), the element showing the second largest quantity among the critical rows of the given D31' matrix Select and use it as a critical element in 5step3
Repeat the following once again. Then, the 10th row is set as the Dn boundary row, and the (corrected) fundamental matrix p%++ that satisfies the critical condition
and a residual matrix p%+3 is obtained.

(step) Next, D2″ = Dfl′″1)
As a result, the operation is repeated once from 5tcpl until D2(1)=O.

Finally, the traffic demand matrix is expanded into a collection of basic matrices as shown below.

p = D, ill + DIZl + p, (
31+...+D-Katana) Here, p, (1' (where I=1. 2...β) is the connection mode between beams (time slot (allocation).

However, if it is attempted to use repeaters connected to each incoming line and outgoing line of a switch with 100% efficiency, a plurality of critical rows (columns) in the traffic demand matrix often occur.

In this case, in the conventional extraction method described above, [5 step 2
) and in [5step 3], elements are selected without a method to find a combination of elements that can extract non-zero elements (elements whose fireflies are not zero) from all critical rows (columns) (step 9). is repeated forever,
A demand assignment service that requires determining the switch switching pattern in real time or in a relatively short period of time in a switching system using space-divided switches has the disadvantage that there are frequent cases where the switch switching pattern cannot be determined. Alternatively, this has become a serious obstacle in realizing a time-dependent reassignment service that requires changing the switching pattern according to traffic distribution at a predetermined time.

Note that, depending on the conventional extraction method described above, (ste
A case in which a switch switching pattern cannot be determined because p 9 ) is repeated forever will be specifically explained below using an example. FIG. 9 is an explanatory diagram for this purpose.

Now, consider a traffic demand matrix as shown in FIG. 9(a). Let the critical row of this matrix be the fourth row (row sum is (
5+5=10).

When the above-mentioned extraction algorithm is applied to this matrix, first, the maximum element of the critical row d ij ” d 43
=5 is extracted. Next, by extracting elements from the matrix other than the fourth row and third column in descending order, by simply scanning from left to right and top to bottom, d, -5 is extracted, and in the same way. dzz=5. dGo4=0 is extracted. Basic matrix DI composed of extracted elements
The residual matrix D2 composed of the remaining elements is as shown in FIG. 9 (b).

However, in D2 of FIG. 9(b), the fourth row, which was originally a critical row, does not satisfy the R boundary condition (row sum is maximum), so the element d in the critical row of Dl.

=5 is returned to the corresponding element of D2. Then, since the critical row of DI is not satisfied, all remaining elements of Dl are D
Returned to 2. In other words, as shown in Figure 9 (c),
The situation is the same as if no extraction had been performed.

Next, according to (step 9), the critical row (fourth row),
Let the second largest element from 2 to 2 be the TJ'B field element. In other words, d4+=5 is extracted from the matrix shown in FIG. 9(a). As for the remaining elements, by simple scanning as before, dz=
5. dzz=5. Then, dss=o is extracted.

In other words, the fundamental matrix and the residual matrix D2 as shown in FIG. 9(d) are obtained.

In this case as well, when the critical condition is taken into account, the result is as shown in FIG. 9 (e), and nothing is extracted from D, and D returns to its original matrix.

As can be seen from this example, when applying the conventional sampling algorithm to a certain kind of demand matrix, (st
ep9) is repeated forever and cannot generate a solution for time slot assignment.

[Purpose of the invention]

The present invention has been made to improve the drawbacks of the prior art as described above, and therefore, an object of the present invention is to solve the problem even when a given traffic demand matrix has a plurality of critical rows (columns). , the demand matrix can always be expanded into a collection of basic matrices containing at most one non-zero element in each row and column, without falling into endless repetition of the same steps,
An object of the present invention is to provide a switching control method for a space-division switch that enables determination of an optimal switch switching pattern.

[Key points of the invention]

The switching control method of the space division type switch according to the present invention has a plurality of input terminals (X1, X2,...Xn) and a plurality of output terminals (Y1, Y2...Yn), and the input terminal and output In a space-division switch that switches between terminals, each traffic that attempts to go from each of the input terminals (X1, X2,...Xn) to each of the plurality of output terminals (Y1, Y2,...Yn) 1) A first means for creating a traffic demand matrix D51' consisting of each traffic amount as an element on the child side; A collection of basic matrices containing at most one zero element (D, (1) + D, +21 +・
. and in the second means, when obtaining the first basic matrix D1(1) from the traffic demand matrix D(1), (a) for each row (column) in the matrix D(1), The sum of the quantities indicated by each element belonging to each row (column) is
When comparing each row (column) unit, the maximum row (column) is considered the critical row (column). There may be more than one critical row), and the maximum amount of elements in each critical row. Select the element (maximum element) that represents the maximum amount, return the element that represents the minimum amount among these maximum elements and set it as dM, extract each element with an amount greater than or equal to dM from the matrix D(1), and calculate the basic An element (hereinafter referred to as an extractable element) that can be taken into the matrix D1(1) and used as its element is defined, and (Tl
) If there are multiple elements that can be extracted, it is decided to select any one of them and incorporate it into the fundamental matrix p,+11, and then the row (column) to which the selected element belongs is The other extractable elements included are taken as index elements, the extractable elements are selected from the column (row) to which the index element belongs, and the basic matrix D is created.
1 (11), (c) However, the critical row containing only one extractable element (
column), it is decided to preferentially select the extractable elements included in the critical row (column) and incorporate them into the fundamental matrix DI'', and at least the steps (a) to (c) above are included. The first fundamental matrix D1(+1) is determined according to the algorithm. After that, the residual matrix of the traffic demand matrix D(1) for the fundamental matrix pl(+1) is set as the traffic demand matrix D321, and the same procedure is repeated to determine the second The key point is that the fundamental matrix DI'z' is obtained for the first time, and the same procedure is repeated thereafter.

[Embodiments of the invention]

The present invention will now be described in detail with reference to the drawings.

FIG. 1 shows a bronc concave illustrating an embodiment of the present invention. In the figure, 1 is a call information input terminal, 2 is a call information management unit,
3 is a traffic demand matrix creation section, 4 is a switch switching pattern setting section, 5 is a switch switching pattern memory section,
6 is a switch switching control section, 7 is a space division type switch section,
8 is a data burst (traffic) input terminal, 9 is a receiver, 10 is a transmitter, and 11 is a data burst (traffic) output terminal.

The outline of the operation is as follows. Based on the call information from each station that is input from the call information input terminal 1, the call information management section 2 collects traffic information for each area, and based on the traffic information, the traffic demand matrix creation section 3 collects the traffic information for each area. , create a traffic demand matrix between each area.

Based on the traffic demand matrix, the switch switching pattern setting unit 4 follows the flow chart shown in FIG. 2 as the flow of the operation of the switch switching pattern setting unit 4 (details will be described later)!・Expand the Rahisoku demand matrix into a collection of basic matrices (matrix containing at most one non-zero element in each row and column) as shown below.

D”D+ (1) + D, (21+ D,
(K) + ・・・・・・+ p, (+1 +
... + p, (blade) here D%I) (where I = 1.2...l) each corresponds to a partial switch switching pattern, so D I ” '〜D
1(λ) is stored in the switch switching pattern memory section 5. Based on the memory pattern, the switch changeover control section 6 controls the space division type switch section 7 in order to switch the received data burst (traffic) from the receiver 9 to the corresponding route. Output to lO.

If the transceiver 9.10 (repeater) is to be used efficiently, a plurality of customer view rows (columns) will exist in the traffic demand matrix, as described above. Even when such a traffic demand matrix is given, the present invention can efficiently determine a switch switching pattern that minimizes the frame length and the number of switching operations, thereby achieving efficient switch switching. realizable. Further, since the number of partial patterns constituting the switch switching pattern is also minimized, the amount of memory for storing the switching patterns can be reduced.

Next, the operation in the switch switching pattern setting section 4 (ie, the procedure for expanding a given traffic demand matrix into a collection of basic matrices) will be explained in detail.

(Step 1) Among the given traffic demand matrix D(1), the one in which the sum of the amounts indicated by each element in each row (or column), that is, the row (column) sum is the largest [for example, K
. Find the first row (column)] at − and call it the critical row (column).

(Step 2) Find the element with the minimum amount among the elements (maximum elements) indicating the maximum amount in each critical row (column), find the value dM of the minimum amount, and calculate the value dM for each critical row (column).
A critical matrix DC'' is created by extracting elements with types greater than or equal to dM as extractable elements.

(3tep 3) If there is a row (column) whose extractable element is 1 among the critical matrix DC (I+, its critical row (column) (KO-th row (column))), then Its extractable element in row (column) is element dB(1,j)
Let this be the basic matrix, +11 elements d.

If it is extracted as (1, j), go to step 5), otherwise go to (step 4).

(Step 4) Any extractable element in the critical matrix Dc'11 is first set as the element d, (t, J) to be extracted, and if this element ds(i, j) is not 0 (zero), then its ds( 1, J) as the element d of the fundamental matrix p, +1>.

Extract as (1, j). If the element ds (1, j) is 0, the maximum element among the elements in the rows and columns that have not yet been extracted in the critical matrix Dc (I) is the element d+( 1. Extract as J).

(step 5) If non-zero elements are extracted from all critical rows (columns) in critical matrix DC(1) (step 7)
If not, proceed to (steρ6).

(Step 6) In e-boundary matrix DC (11), the row belonging to the element extracted as element d1 (1, j) (
When there is a non-zero element as an extractable element in a column (column), the non-zero element is used as an index element, and the largest element in the column (row) to which the index element belongs is the next element to be extracted di( 1
, j). If the element does not exist, di (t
, j)=O. After that, return to step 3.

(Step 7) Given traffic demand matrix D'
'', elements of the basic matrix D, '' are repeatedly extracted from the rows (columns) other than the critical rows (columns) in order of the large loss, and when the residual matrix is set as D2(1), D(column)
=D, (11+1) 2 (expands into the shape of an eye.

(Step 8) Critical rows (columns) of fundamental matrix D1'', residual matrix D2'' and given traffic demand matrix D(1)
The critical row (column) is guaranteed so that the critical row (column) of is defeated ((step 7) to (step 8) of the prior art)
reference).

(Step 9) Next, D (+41+ = p%1
1, a number of times (s
Repeating step 1), the traffic demand matrix is expanded into a collection of basic matrices as shown below.

D = p, fl + D, +21 + p,
(:11 + ・・・・・・+ D, (1) +・
...P D, (J) As mentioned above, the procedure for expanding a given traffic demand matrix into a collection of basic matrices is shown in the flowchart in Figure 2, as described above. Since they are shown together, it will be easier to understand by referring to FIG. 2 as well.

Next, a specific example of its application will be explained with reference to FIG.

The given traffic demand matrix D(1) is shown in Fig. 3(a
Suppose that it is as shown in l. In this case, the arrow (i=2
．． 3. 4, j=2.3) is the critical row (column).

Next, from (step 2), since the maximum element in all critical rows (columns) is 2, dM=2 after all, and when the boundary matrix D%I+ is created, it becomes as shown in (b).

In (step 3) 1) n-boundary row (column) Dc'I+
Since there is no row (column) in which the number of elements is 1, (s
Step 4), and find the maximum element 2 (i=4゜j=3), which is one of the extractable elements in the critical row (column) pc+11.
is the fundamental matrix and extracted as +11 elements.

Next, from (step Ei), the element that becomes the index for extraction of the next element ds (+, J) to be extracted in the same row (column) as 2 (i=4.j=3) is 2 (i=4.j=3). j=1) and 2(i
=3. j=3), select one of them, for example, 2 (i=3.j=3), and select the largest element in the column to which this element belongs, that is, 2 (i=3. j=2) is the element dS(iIJ) to be extracted next.

Next, return to (step 4) via (step 3) d.

Extract (1, j) = ds (1, j) = 2 (3, 2),
(Step 6) d using 2 (2, 2) as an index.

Choose (1, j)=2(2,1).

Next, through (step 3), the process returns to (step 4), where d+(1, j)=ds(1, j)=2(2
, 1), and in (step 5), the critical matrix DC
Since non-zero elements are extracted from all critical rows in (1), the fundamental matrix DI before guaranteeing critical rows (columns)
” and the residual matrix D2(1)” are shown in Figure 3 (c 1), (c
2).

(step 8), the given traffic demand matrix D(1
) is included in each critical row of Dl(1)′ and D2(n′), so the fundamental matrix D1(1)=D
l(1)', and the residual matrix D2(1)=D2''1'.

Next, as D0ゝ = D2(I), that is, the residual matrix D
2(1) as the given traffic demand matrix D32)
Then, in (step 2), since dM=1, p ci21 = D +! 1 In other words, the critical matrix D C'' and the traffic demand matrix p(Z) are equal, and by the same procedure, 2 (3, 3), 1
(4, 4), 2 (2゜2) are extracted in the order of the second fundamental matrix Dρ2)' before guaranteeing the critical row (column),
The residual matrices D2(2)' are shown in Figure 3 (d 1), (d 2
)become that way.

Coco T: D + ” ' + Dz ” ” '
i: H'A is not guaranteed, so the corresponding elements of p, (2 (2゜2) of Zl' and 1 each from 2 (3, 3) D%Z)' If we return to the equation and guarantee the critical row (column), we get the result as shown in FIG. 3 (e 1) + (e 2).

Below, the third fundamental matrix D I ”l residual matrix D 2
” and the fundamental matrix D I 'l of the fourth round, and the residual matrix D2 (4+).
), (g1).

(g2), and in the end pf+) -ri, N)
+D, (Z) + p, fil + p
, (41), thereby creating a switch switching pattern.

Next, creation of a switching pattern will be explained.

The fundamental matrix DI"' is shown in Figure 3 (C1), D1'21 is shown in Figure 3 (el), D, (31 is shown in Figure 3 (fl), D
l'4) are shown respectively in FIG. 3 (gl), so FIG. 4 shows them together.

A switch switching pattern created using the four basic matrices shown in FIG. 4 is shown in FIG.

The pattern shown in the figure has the minimum frame length and the minimum number of switch changes.

〔Effect of the invention〕

As explained above, according to the present invention, a switch switching pattern that minimizes the frame length and the number of switching operations is determined and determined even if any traffic demand matrix based on various call information is given. It is possible to control the switching of the space-division type switch by Furthermore, there is an advantage that the capacity of memory for storing switch switching patterns can be reduced.

Especially suitable for the SS-TDMA method in satellite communication,
When the present invention is applied to an IF switch that switches TDMA signals in the microwave band, transponders must be used 100% of the time in order to make effective use of the satellite link, but as a result, the traffic Since most rows and columns of the demand matrix are critical rows (columns), it is believed that the present invention is effective in determining the switching pattern on the premise that there are a plurality of critical rows (columns) in the traffic demand matrix. I can say that.

Furthermore, a group-based switching system using space-division switches, which is considered essential for the realization of variable relay networks in the future (Shimazaki, "Proposal of a group relay switching system", IEICE Technical Report, SE 7)
4-63, pp. 81-93, 1974).

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a flow chart showing the operation flow of the switch switching pattern setting unit 4 in FIG. FIG. 4 is an explanatory diagram of a specific application example of the procedure used in the present invention to expand into a collection of
Fig. 5 is an explanatory diagram of the switch switching pattern created using the four basic matrices shown in Fig. 4, and Fig. 6 is an explanatory diagram of the switch switching pattern created using the four basic matrices shown in Fig. 4.
A conceptual diagram of the MA system, Fig. 7 is an explanatory diagram of the traffic demand matrix, Fig. 8 is an explanatory diagram of the switch switching pattern, and Fig. 9 is an explanatory diagram of the traffic demand matrix.
The figure is an explanatory diagram specifically showing the drawbacks of the extraction method according to the prior art. Code explanation 1...Call information input terminal, 2...Call information management section, 3.
...Traffic demand matrix creation unit, 4...Switch switching pattern setting unit, 5...Switch switching pattern memory unit, 6...Switch switching control unit, 7...Space division type switch unit, 8...・Data burst input terminal, 9
...Receiver, 10...Transmitter, 11...Data burst output terminal, 100...Communication satellite equipped with space division switch, Ai...Spot beam area covered by the same frequency, Xi... ...uplink, Yi...downlink. Agent Patent Attorney Akio Namiki Agent Patent Attorney Matsuzaki Kyouzuki T6 WE3 Figure 4 Figure 5 Figure 1 Frame 6 Figure 7 Figure aS 1 Frame

Claims (1)

[Claims] 1) It has a plurality of input terminals (X1, X2,...Xn) and a plurality of output terminals (Y1, Y2,...Yn), and switches between the input terminals and the output terminals. In the space-division type switch, for each amount of traffic going from each of the input terminals (X1, X2,...Xn) to each of the plurality of output terminals (Y1, Y2,...Yn), for each input terminal, A first means for creating a traffic demand matrix D^(^1^) consisting of each traffic amount as an element for each output terminal; and a non-zero element in each row and each column of the matrix D^(^1^). A collection of basic matrices containing at most one (D1^(^1^)+D1^
(^2^)+...D1^(^n^)), and creating a switching pattern of the switch from the collection of the obtained basic matrices, and controlling switching of the switch according to the pattern. and a third means for performing (I ) For each row (column) in the matrix D^(^1^), the row (column) in which the sum of the quantities indicated by each element belonging to each row (column) is the largest when compared for each row (column). is a critical row (column) (there may be more than one critical row), select the element (maximum element) that shows the largest amount among the elements in each critical row, and then Select the element that shows the minimum amount and set it as dM, and extract the elements with the amount greater than or equal to the dM from the matrix D^(^1^), incorporate them into the basic matrix D1^(^1^), and calculate that element. (hereinafter referred to as extractable elements), and (b) If there are multiple extractable elements, select any one of them and create the basic matrix D1^(
^1^), and then set other extractable elements included in the row (column) to which the selected element belongs as an index element, and extractable elements from the column (row) to which the index element belongs. , and decide to incorporate it into the fundamental matrix D1^(^1^), (c) However, the critical row containing only one extractable element (
column), it is decided to preferentially select the extractable elements included in the critical row (column) and incorporate them into the basic matrix D1^(^1^), and at least the above (a) to (c) Determine the first fundamental matrix D1^(^1^) according to the algorithm including the steps of
After that, the residual matrix of the traffic demand matrix D^(^1^) with respect to the basic matrix D1^(^1^) is set as the traffic demand matrix D^(^2^), and the same procedure is repeated to obtain the second matrix. A switching control method for a space division type switch, characterized in that the basic matrix D1^(^2^) is obtained for the first time, and the same procedure is repeated thereafter.