1 5 3 x11 10
2
7 4 x 21 15
Abstract—This is the first series of research papers to define
multidimensional matrix mathematics, which includes
multidimensional matrix algebra and multidimensional matrix 52 8 x 31 7
calculus. These are new branches of math created by the author * =
with numerous applications in engineering, math, natural 10 4 3 x12 13
8
science, social science, and other fields. Cartesian and general 5 1 x 22 18
tensors can be represented as multidimensional matrices or vice 2 6 9 x 32 19
versa. Some Cartesian and general tensor operations can be
performed as multidimensional matrix operations or vice versa. Any of the different methods used to solve systems of linear
However, many aspects of multidimensional matrix math and equations represented with classical matrices can be applied
tensor analysis are not interchangeable. Part 6 of 6 describes to each of the submatrices in multidimensional matrices. This
the solution of systems of linear equations using includes graphing, the substitution method, the elimination
multidimensional matrices. Also, the multidimensional matrix
method, Gaussian elimination, Gauss-Jordan elimination,
calculus operations for differentiation and integration are
defined. Cramer’s rule, LU decomposition, Cholesky decomposition,
etc.
Index Terms—multidimensional matrix math, The preceding system of linear equations can be
multidimensional matrix algebra, multidimensional matrix represented with the following augmented 3-D matrix:
calculus, matrix math, matrix algebra, matrix calculus, tensor 1 5 3 10
analysis
2 7 4 15
5 2 8 7
I. INTRODUCTION 10 4 3 13
Part 6 of 6 describes the solution of systems of linear
equations using multidimensional matrices. Also, part 6 of 6 8 5 1 18
2 6 9 19
defines the multidimensional matrix calculus operations for
differentiation and integration. Using Gauss-Jordan elimination on each of the individual
2-D submatrices in the augmented 3-D matrix, the reduced
row echelon form for this augmented multidimensional matrix
II. SOLVING SYSTEMS OF LINEAR EQUATIONS WITH can be found:
MULTIDIMENSIONAL MATRICES
1 5 3 10 1 0 0 37 25
Systems of linear equations can be represented and solved
with multidimensional matrices. 2 7 4 15 0 1 0 51 25
5 2 8 7 0 0 1 14 25
Consider the following system of linear equations:
~
x11 + 5x21 + 3x31 = 10 10 4 3 13 1 0 0 13 112
2x11 + 7x21 + 4x31 = 15
5x11 + 2x21 + 8x31 = 7 8 5 1 18 0 1 0 31 8
10x12 + 4x22 + 3x32 = 13 2 6 9 19 0 0 1 25 56
8x12 + 5x22 + x32 = 18
2x12 + 6x22 + 9x32 = 19 37 51 14 13
Therefore, x11 = ; x21 = ; x31 = ; x12 = ; x22
This system of linear equations can be represented with the 25 25 25 112
following multidimensional matrix equation composed of 3-D 31 25
matrices: = ; x32 =
8 56
Consider the following system of linear equations:
a11111x1111 + a12111x2111 + a13111x3111 = b1111
Manuscript received March 23, 2010. a21111x1111 + a22111x2111 + a23111x3111 = b2111
Ashu M. G. Solo is with Maverick Technologies America Inc., Suite 808, a31111x1111 + a32111x2111 + a33111x3111 = b3111
1220 North Market Street, Wilmington, DE 19801 USA (phone: (306) a11121x1121 + a12121x2121 + a13121x3121 = b1121
242-0566; email: amgsolo@mavericktechnologies.us).
a21121x1121 + a22121x2121 + a23121x3121 = b2121
This system of linear equations can be represented with this single multidimensional matrix equation composed of 5-D matrices:
a11111 a12111 a13111 a11121 a12121 a13121 x1111 x1121 b1111 b1121
a 21111 a 22111 a 23111 , a 21121 a 22121 a 23121 x 2111 , x2121 b 2111 , b 2121
a 31111 a32111 a33111 a31121 a32121 a33121 x3111 x3121 b3111 b3121
a11211 a12211 a13211 a11221 a12221 a13221 x1211 x1221 b1211 b1221
a 21211 a 22211 a 23211 , a 21221 a 22221 a 23221 x 2211 , x2221 b 2211 , b 2221
a 31211 a32211 a33211 a 31221 a32221 a33221 x3211 x3221 b3211 b3221
*
=
a11112 a12112 a13112 a11122 a12122 a13122 x1112 x1122 b1112 b1122
a 21112 a 22112 a 23112 , a 21122 a 22122 a 23122 x 2112 , x2122 b 2112 , b 2122
a 31112 a32112 a33112 a31122 a32122 a33122 x3112 x3122 b3112 b3122
a11212 a12212 a13212 a11222 a12222 a13222 x1212 x1222 b1212 b1222
a 21212 a 22212 a 23212 , a 21222 a 22222 a 23222 x 2212 , x 2222 b 2212 , b 2222
a 31212 a32212 a33212 a31222 a32222 a33222 x3212 x3222 b3212 b3222
When each element aijklm of the coefficient matrix and each multidimensional matrix has the same number of dimensions
element bjklm of the product matrix is defined, each element and same number of elements in each dimension as the
xjklm of the variable matrix can be calculated. Any of the multidimensional matrix that is integrated.
different methods used to solve systems of linear equations A 3-D matrix with dimensions of 2 * 2 * 2 is integrated as
represented with classical matrices can be applied to each of follows:
the submatrices in multidimensional matrices. The system of 0 sin x K cos x
linear equations can be solved like in the preceding example. xy
2
6 x ye 3 x e xy
e
x
2y
dx =
e
x
2 xy
III. MULTIDIMENSIONAL MATRIX DIFFERENTIATION sec x tan x
sec x tan x sec x
2
In multidimensional matrix calculus, multidimensional
matrices are differentiated by finding the derivative of each
element in the multidimensional matrix. The result is a
multidimensional matrix of derivatives. The resulting V. CONCLUSION
multidimensional matrix has the same number of dimensions Part 6 of 6 described the solution of systems of linear
and same number of elements in each dimension as the equations using multidimensional matrices. Also, part 6 of 6
multidimensional matrix that is differentiated. defined the multidimensional matrix calculus operations for
A 3-D matrix with dimensions of 2 * 2 * 2 is differentiated differentiation and integration.
as follows: Classical matrix math offers many benefits not present in
5 cos x 0 sin x tensor analysis for a first or second order tensor, and tensor
2 analysis for a first or second order tensor offers many benefits
e x 6 x e
x
d 3 x not present in classical matrix math. Similarly,
=
e 2
dx e
2x
x x multidimensional matrix math offers many benefits not
sec x tan x present in tensor analysis for tensors of any order, and tensor
sec x tan x sec x
2
analysis for tensors of any order offers many benefits not
present in multidimensional matrix math.
The author predicts that multidimensional matrix math will
IV. MULTIDIMENSIONAL MATRIX INTEGRATION replace classical matrix math in the future when this subject is
In multidimensional matrix calculus, multidimensional taught in university courses. The author has made many more
matrices are integrated by finding the integral of each element developments in multidimensional matrix math and
in the multidimensional matrix. The result is a developed many more innovative applications of
multidimensional matrix of integrals. The resulting multidimensional matrix math that will soon be published.
REFERENCES
[1] Franklin, Joel L. [2000] Matrix Theory. Mineola, N.Y.: Dover.
[2] Young, Eutiquio C. [1992] Vector and Tensor Analysis. 2d ed. Boca
Raton, Fla.: CRC.