Scribd
Upload
59 views

X T X So That The Coe Cients of X: Step 2. Transform

This document describes an extended algorithm for solving underdefined multivariate quadratic equations (MQ). The algorithm transforms the equations into a standard form through linear transformations of the variables. It then solves systems of linear equations to obtain quadratic equations in isolated variables, which can be solved to find the values of those variables. The complexity of this algorithm is O(nw m(log q)2) when the field characteristic is 2, and O(2m nw m(log q)2) when the characteristic is odd, where w is the exponent of Gaussian elimination, n is the number of variables, m is the number of equations, and q is the size of the finite field.

Uploaded by

mipouyg
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
59 views

X T X So That The Coe Cients of X: Step 2. Transform

This document describes an extended algorithm for solving underdefined multivariate quadratic equations (MQ). The algorithm transforms the equations into a standard form through linear transformations of the variables. It then solves systems of linear equations to obtain quadratic equations in isolated variables, which can be solved to find the values of those variables. The complexity of this algorithm is O(nw m(log q)2) when the field characteristic is 2, and O(2m nw m(log q)2) when the characteristic is odd, where w is the exponent of Gaussian elimination, n is the number of variables, m is the number of equations, and q is the size of the finite field.

Uploaded by

mipouyg
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

Extended Algorithm for Solving Underdefined MQ 121

Step 2. Transform x → T3 x so that the coefficients of x1 x3 , x2 x3 in fj (j =


1, . . . , m) are zero.
⎛ ⎞
⎛ ⎞ ∗ 0 0
∗ 0 ⎜ ⎟
⎝ 0 ∗ ⎠ → ⎜ 0 ∗ 0 ⎟
∗
⎝ 0 0 ∗ ⎠
∗
..
.
(We continue similar operations to “Step m − 1.”.)
From “Step 1.” to “Step m − 1.”, we require the condition n−1 ≥ m(m−1),
i.e., n ≥ m2 − m + 1.
Then we can obtain the coefficient matrices of the form
⎛ ⎞
∗ 0
⎜ .. ⎟
⎜ . ⎟
⎜ ⎟
⎜ 0 ∗ ⎟
⎜ ⎟
⎜ ⎟
⎜
⎝
⎟
⎠ ∗
for each i = 1, . . . , m, and the following quadratic equations.
⎧ m
⎪ m
⎪
⎪ β x 2
+ xi L1,i (xm+1 , . . . , xn ) + Q1 (xm+1 , . . . , xn ) = 0
⎪
⎪
1,i i
⎨ i=1 i=1
.. (3)
⎪ .
⎪
⎪
⎪
⎪
m
2
m
⎩ βm,i xi + xi Lm,i (xm+1 , . . . , xn ) + Qm (xm+1 , . . . , xn ) = 0
i=1 i=1

where L’s are linear polynomials and Q’s are quadratic polynomials in these
variables.
Step m. Solve linear equations {Li,j (xm+1 , . . . , xn ) = 0} for i = 1, . . . , m, and
j = 1, . . . , m, and substitute the solutions xm+1 , . . . , xn into (3). This system
of linear equations has n − m unknowns and m2 equations, so we can solve if
n and m satisfy n − m ≥ m2 i.e. n ≥ m(m + 1). Finally, we obtain quadratic
equations of the form (2). Then we can compute the x21 , . . . , x2m values easily.
The complexity of this algorithm is

O(nw m(log q)2 ) (char k is 2)
O(2m nw m(log q)2 ) (char k is odd),
where 2 ≤ w ≤ 3 is the exponent of the Gaussian elimination. This is because this
algorithm computes n × n matrices over finite field k = GF(q) and solves linear
equations to obtain the x21 , . . . , x2m values. The complexity of these operations is
O(nw (log q)2 ). When the characteristic of k is odd, the probability of existence
of square roots is approximately 1/2, and we can find a solution with probability
of 2−m . Therefore, when the characteristic of k is odd, the complexity of this
algorithm is O(2m nw (log q)2 ).

You might also like