TEKLA MUMBI BITC01/2071/2022

EMMANUEL MUTAI BITC01/2155/2022

ALGEBRAIC ALGORITHMS

Introduction.
Algebraic algorithms come into play when dealing with problems that have a natural
mathematical representation. By expressing the problem and potential solutions algebraically,
we can design algorithms with strong theoretical foundations and analyze their efficiency using
mathematical tools. It deals with numbers, vectors, matrices, polynomials, formal power series,
Exponential and different polynomials, rational functions, algebraic sets, curves and surfaces.
Matrix computations and Approximation of polynomial zeros

Algebraic algorithms excel at manipulating matrices, essential for tasks like solving systems of linear
equations, performing matrix multiplication, and finding eigenvalues. These operations form the
backbone of many scientific computing and graphics applications.

This section also covers several algebraic and numerical problems of computing that are usually
solved numerically by rounding off or chopping the input and computed values to a fixed number
of bits that fit the computer precision.

Product of vectors and matrices

In linear algebra, the product of a matrix and a vector is a fundamental operation. However, there's
a specific compatibility requirement. The number of columns in the matrix must equal the number
of entries (or rows if the vector is written as a row vector) in the vector. When this condition is met,
the resulting product is another vector. The calculation involves multiplying each element in a row
of the matrix with the corresponding element in the vector and summing the products. Essentially,
it's a weighted linear combination of the matrix's columns, where the weights are provided by the
vector's entries. This operation is crucial in various applications, from solving systems of linear
equations to transforming points in computer graphics

Example:

Consider the following:

 Matrix A: | 1 2 | | 3 4 |
 Vector x: ( -1 ) ( 2 )

Here, A is a 2x2 matrix (2 rows, 2 columns) and x is a 2x1 vector (2 rows, 1 column). Since the
number of columns in A (2) matches the number of entries in x (2), we can perform the
product.

To calculate Ax (the product of A and x):

1. Row-by-column multiplication: For each row of A, multiply the corresponding elements with the
vector x and sum the products.
o First row of A: (1 * -1) + (2 * 2) = 3
o Second row of A: (3 * -1) + (4 * 2) = 5
2. Resulting vector: Stack the products from each row to form a new vector.

Therefore, Ax = (3) (5)

Gaussian Elimination of Algorithms.

Gaussian Elimination, also known as row reduction, is a powerful algorithmic technique for
solving systems of linear equations. It operates on the coefficient matrix that represents the
system, transforming it through a series of well-defined row operations. These operations don't
change the underlying solutions to the system. The algorithm strategically manipulates the
rows to achieve a specific structure called echelon form, where zeros strategically appear below
the diagonal entries. This form allows for a straightforward back-substitution process to
efficiently solve for the unknown variables one by one. Gaussian Elimination offers a systematic
and efficient approach to solving linear systems, making it a cornerstone of many scientific
computing applications and forming the foundation for more advanced algorithms in linear
algebra.

Example

Equations:

 x1 + 2x2 + 3x3 = 14 (Equation 1)

 2x1 + 5x2 + x3 = 7 (Equation 2)
 x1 + 3x2 - 2x3 = -1 (Equation 3)

Step 1: Write the Augmented Matrix

First, we convert the system of equations into an augmented matrix, where each row
represents one equation. The coefficients of the variables go in the first columns, and the
constants on the right side go in the last column.

| 1 2 3 | 14 |
 |2 5 1| 7|
| 1 3 -2 | -1 |

Step 2: Eliminate x1 from Equation 2 and 3 (using Equation 1)

We want to get zeros below the x1 term in the first column (except for the first row itself).

 In Equation 2, multiply Equation 1 by -2 and add it to Equation 2.

 In Equation 3, multiply Equation 1 by -1 and add it to Equation 3.

The new augmented matrix becomes:

| 1 2 3 | 14 |
| 0 1 -5 | -5 |
| 0 1 -5 | -15 |

Step 3: Eliminate x2 from Equation 3 (using Equation 2)

Similar to step 2, we want a zero below the x2 term in the second column (except for the
second row).

 In Equation 3, add Equation 2 to Equation 3.

The new augmented matrix becomes:

| 1 2 3 | 14 |
| 0 1 -5 | -5 |
| 0 0 -10 | -10 |

Step 4: Back-Substitution

Now the matrix is in upper triangular form. We can solve for the variables using back-
substitution:

1. Solve the last equation for x3: x3 = 1

2. Substitute x3 back into the second equation to solve for x2: x2 = 0
3. Substitute x3 and x2 back into the first equation to solve for x1: x1 = 10

Solution:

Therefore, the solution to the system of equations is:

X1 = 10, x2 = 0, x3 = 1
 :

Polynomial factorization.

Polynomial factorization is the process of breaking down a polynomial expression into a product of
simpler polynomial expressions, much like factoring integers into prime numbers. These simpler
factors, called irreducible polynomials, cannot be further factored using polynomial operations.
Factorization is a crucial concept in algebra, with applications in various areas of mathematics and
beyond.

Example:

Let's factor the polynomial expression: x^2 + 5x + 6

Here, we're looking for two linear expressions (of the form ax + b) that multiply to give the
original polynomial. We can try to find two values, a and b, such that:

 (ax + b) * (cx + d) = x^2 + 5x + 6

By expanding the product on the left and comparing coefficients with the original expression,
we can solve for a, b, c, and d. In this case, we find that a = c = 1, and b = 2, d = 3.

Therefore, the factored form of the polynomial is:

(x + 2) (x + 3)

Applications of algebraic algorithms in design and analysis of algorithms.

 Number Theoretic Algorithms: Integer factorization, primarily testing, and modular arithmetic
all rely heavily on algebraic algorithms. These algorithms are crucial for cryptography, where
secure communication hinges on the difficulty of factoring large numbers.

 Fast Fourier Transforms (FFTs): These efficient algorithms for multiplying polynomials have
revolutionized signal processing and computer graphics. They exploit the properties of
polynomial rings and roots of unity to achieve significant speedups compared to naive
multiplication.

 Geometric Algorithms: Many geometric problems, such as finding intersections of lines or

determining convex hulls, can be formulated algebraically. By leveraging properties of points,
lines, and planes in coordinate systems, algebraic algorithms provide efficient solutions for
geometric computations.

 Symbolic Computation: This field deals with manipulating expressions symbolically rather than
numerically. Algebraic algorithms form the foundation for symbolic manipulation systems,
 allowing for tasks like simplifying expressions, solving polynomial equations symbolically, and
performing automated differentiation (finding the derivative of a function).

 Complexity Analysis: Understanding the behavior of algorithms often involves analyzing their
running time or space complexity. By expressing algorithms algebraically and analyzing the
growth rate of involved functions (like polynomials), we can determine their theoretical
complexity and compare different approaches.

The Sylvester resultant system of nonlinear equation

The Sylvester resultant is a tool used to analyze a system of polynomial equations. It helps
eliminate one of the variables and convert the system into a single polynomial equation in
another variable.

Here is a breakdown of the concept:

System of Polynomial Equations: We consider a system of polynomial equations in two

variables, say x and y. These equations can be linear (containing only terms with degree 1 in x
and y) or nonlinear (containing terms with higher degrees).

Sylvester Resultant: This method constructs a new polynomial equation, called the resultant
that depends only on the coefficients of the original system. The resultant has certain
properties:

Existence of Solutions: If the resultant is non-zero, then the original system has a finite number
of solutions (x, y). However, a zero resultant does not necessarily imply no solutions (it could
indicate infinitely many solutions or none at all).

Degree of Resultant: The degree of the resultant (the highest power of the variable) is equal to
the product of the degrees of the original polynomials in the eliminated variable.

Grobner bases

Grobner bases are a powerful tool in abstract algebra, particularly in computer algebra and
related fields. They help analyze and solve problems involving systems of polynomial equations,
especially nonlinear ones.

Gröbner bases do not directly give solutions, but they provide a simplified way to analyze and
solve the system.

The specific form of the Gröbner basis depends on the chosen monomial ordering, but it will
always capture the essential information about the original system.
 Conclusion.

In conclusion, algebraic algorithms play a vital role in the design and analysis of algorithms.
They offer a powerful framework for tackling problems with a natural mathematical structure,
from cryptography and signal processing to geometric computations and symbolic
manipulation. By leveraging the elegance and precision of algebra, we can design efficient and
provably correct algorithms, analyze their complexity, and unlock new possibilities in the ever-
evolving world of computer science. As the field continues to develop, we can expect even
more innovative applications of algebraic algorithms to emerge, shaping the future of efficient
and reliable computation.