The partial Bell Polynomial expresses polynomials of any order and is convenient to calculate parent sequences and convolutions. The Perrin sequence is used as an example showing that a convolution hierarchy of sequences of various order can be algebraically equal. The condition for algebraic equality is shown and how the calculus (derivatives) of these convolutions are determined. Linear mappings of the Parent and Element sequences are shown to be group homomorphisms.
Algebra of Integer sequences and their Convolution using Partial Bell Polynomials
The partial Bell Polynomial is modified and shown to represent many integer sequences in the literature. The modified Bell Polynomial and its derivative are equivalent to Inter Sequence Polynomials and are applied to calculate Perrin and Padovan numbers, respectively. A detailed investigation of how these polynomials can reproduce the integer sequence is shown. The Bell Polynomial and number are important in combinatorics, partitions and mappings. This ability to describe various mappings with higher Bell Numbers provides a better understanding of the relation of integer sequences to discrete dynamic processes.
Bell Polynomials and Perrin and Padovan Sequences
A comparison of integer linear recurrent sequences and the parent and element sequences generated from inter sequence polynomials is presented. Bell Polynomials are shown to be connected to inter sequence polynomials. Bell Polynomials can be used to calculate ISPs of any order through a simple equation.
ISP_LRS and Bell Polynomials
The convolution of sequences using ISPs and their derivatives were described in a previous Chapter (see Chapter 36). The Convolution Recurrences from Inter Sequence Polynomials (CRISP) program is designed to find the coefficients of second order and third order convolutions. The order of these sequences is found to be the sum of the orders or the component sequences. By using the ISPS it was possible to find closed forms of the recurrence coefficients of the second and third derivatives of any linear second and third order parent recurrence. This program (Mathematica) is shown in the attached pdf.
Every linear integer sequence of a cubic polynomial has an associated element sequence. The element sequence is the first derivative of the parent polynomial and easily obtained from its inter sequence polynomial. Some sequences such as the Perrin sequence have an element sequence (the Padovan sequence) which is linearly proportional after rounding to its parent sequence. This chapter discusses a method for finding this associated constant of proportion for any linear integer sequence based on the discriminant. It is found that not all sequences exhibit this proportionality and an attempt is made to find condition for existence.
Discriminants and the Element Sequence
Initiators are a set of numbers which can generate a sequence. This appendix provides simple cubic and quadratic polynomials for calculating the initiator sequences associated with 1st and higher order maximal independent sets. Based on the constant g defined for the hypergeometric equation, sequence values can be readily calculated for odd values of g where g = 3 represents Perrin numbers. The repeated pattern of a(n) = n for small odd n is common to these sequences and suggests a pattern in nature which leads to diversity and variation of structure. An example of a DNA sequence calculated from a higher order MIS sequence is shown (see above) to produce this variation in structure.
An Algorithm for Higher Order Perrin Numbers
This chalkboard discusses the previously derived hypergeometric equation for calculating Perrin numbers. The equation is also found to calculate the number of k-th order maximal independent sets in a cycle graph. Convolution sequences are found to be related to the k-th order numbers and to two combinatorial numbers.
Perrin Numbers, a Hypergeometic Function and Convolution Sequences
This chalkboard demonstrates using ISPs to solve second, third and fourth order homogeneous ordinary differential equations. These solutions agree with the Binet formula solution for second order and Binet-like formula for higher order ODEs. A comparison of these solutions to the Laplace Transform method is also shown. Non-homogeneous ODEs are also found to be solvable by ISP sequences. Equations for finding the solutions to second, third and fourth order non-homogeneous ODEs by ISP are shown.
“You can’t go back and change the beginning, but you can start where you are and change the ending” – C. S. Lewis
This chapter concludes the topic of Inter sequence polynomials. Many polynomial sequences can be defined by ISPs. Among these include Lucas-Fibonacci and Perrin-Padovan polynomial sequences. The derivative of the ISP is used as a solution to second order differential equations.
Higher Order InterSequencePolynomials
A paradox is found between derivatives of the ISPs and discrete convolution of integer sequences. ISP derivatives are convolution operators!