Abstract:Evolving Boolean functions with specific properties is an interesting optimization problem since, depending on the combination of properties and Boolean function size, the problem can range from very simple to (almost) impossible to solve. Moreover, some problems are more interesting as there may be only a few options for generating the required Boolean functions. This paper investigates one such problem: evolving five-valued spectra Boolean functions, which are the functions whose Walsh-Hadamard coefficients can only take five distinct values. We experimented with three solution encodings, two fitness functions, and 12 Boolean function sizes and showed that the tree encoding is superior to other choices, as we can obtain five-valued Boolean functions with high nonlinearity.

What problem does this paper attempt to address?

### What problem does this paper attempt to solve? This paper aims to solve the problem of how to use evolutionary algorithms (EAs) to design five - valued spectra Boolean functions. Specifically, the authors explore how to generate Boolean functions with specific Walsh - Hadamard spectral values through evolutionary algorithms, and these spectral values can only take five different values. #### Background and Motivation 1. **Importance of Boolean Functions**: Boolean functions are of great significance in cryptography, especially when constructing secure encryption systems. Boolean functions with specific properties can provide higher non - linearity and balance, thereby enhancing the security of the system. 2. **Limitations of Existing Research**: Although a large number of studies have shown that evolutionary algorithms can generate Boolean functions with excellent properties, most of the research focuses on problems where known algebraic constructions can achieve excellent or optimal results. Therefore, the practical application value of these new methods is questioned. 3. **Unique Challenges of Five - valued Spectra Boolean Functions**: Five - valued spectra Boolean functions refer to Boolean functions whose Walsh - Hadamard coefficients can only take five different values. Such functions are very interesting in cryptography because they can be balanced and have excellent non - linearity. However, there is less work on the design of five - valued spectra Boolean functions, and there is no research on computational or heuristic methods. #### Research Objectives 1. **Explore New Design Methods**: This paper attempts to use evolutionary algorithms to design five - valued spectra Boolean functions for the first time. This is not only an interesting optimization problem, but also a challenge to the fields of evolutionary algorithms and mathematics. 2. **Evaluate the Effects of Different Encodings and Fitness Functions**: The authors experimented with three solution encodings (including tree encoding, truth - table encoding, and algebraic normal form encoding), two fitness functions, and 12 Boolean function sizes to evaluate which method is more effective. 3. **Explain the Limitations of Traditional Methods**: The paper explains why the commonly used spectral inversion principle cannot be directly applied to the design of five - valued spectra Boolean functions, making this problem more complex. ### Main Contributions 1. **First Application of Evolutionary Algorithms to Design Five - valued Spectra Boolean Functions**: This is the first attempt to use evolutionary algorithms to design such functions, filling the gap in the relevant field. 2. **Discover the Advantages of Tree Encoding**: The experimental results show that for smaller - sized Boolean functions (dimensions 5 and 6), all methods can find five - valued functions well; but for higher dimensions (such as 7 - dimensional and above), only tree encoding can consistently find five - valued functions, and in some odd - sized cases, it reaches the best - known non - linearity. 3. **Explain the Inapplicability of the Spectral Inversion Principle**: It explains in detail why the spectral inversion principle is not applicable to the design of five - valued spectra Boolean functions, providing a theoretical basis for future research. In conclusion, this paper solves the design problem of five - valued spectra Boolean functions by introducing evolutionary algorithms and provides important references and guidance for future research.

The More the Merrier: On Evolving Five-valued Spectra Boolean Functions

A New Angle: On Evolving Rotation Symmetric Boolean Functions

A Systematic Evaluation of Evolving Highly Nonlinear Boolean Functions in Odd Sizes

Look into the Mirror: Evolving Self-Dual Bent Boolean Functions

A Search for Nonlinear Balanced Boolean Functions by Leveraging Phenotypic Properties

Learning Boolean functions with concentrated spectra

Coevolving Boolean and Multi-Valued Regulatory Networks

Generating Bent Functions and Dynamic Filters: A Novel Equivalence-Based Approach

A Discrete Particle Swarm Optimizer for the Design of Cryptographic Boolean Functions

-equivalence classes, a new direction to find better weightwise perfectly balanced functions, and more

Coset leaders of the first-order Reed-Muller codes in four new classes of Boolean functions

Evolving Diploid Boolean and Multi-Valued Gene Networks

Immediate Neighbours of Monotone Boolean Functions

New Geometrical Spectra of Linear Codes with Applications to Performance Analysis

Variable Solution Structure Can Be Helpful in Evolutionary Optimization

Combinatorial spectra using polynomials

The fine structure of the spectral theory on the $S$-spectrum in dimension five

On design-theoretic aspects of Boolean and vectorial bent functions

Differentiability and Optimization of Multiparameter Persistent Homology

On Boolean Functions with Low Polynomial Degree and Higher Order Sensitivity

Strengthening Convex Relaxations of 0/1-Sets Using Boolean Formulas