Teaching the Computer how to Discover(!) and then Prove(!!) (all by Itself(!!!)) Analogs of Collatz's Notorious 3x+1 Conjecture

Doron Zeilberger

DOI: https://doi.org/10.48550/arXiv.0903.4050

2009-03-24

Abstract:Paul Erdos claimed that mathematics is not yet ready to settle the 3x+1 conjecture. I agree, but very soon it will be! With the exponential growth of computer-generated mathematics, we (or rather our silicon brethrern) would have a shot at it. Of course, not by number crunching, but by symbol crunching and automatic deduction. In the present article, I taught my computer how to use the brilliant ideas of four human beings (Amal Amleh, Ed Grove, Candy Kent, and Gerry Ladas) to prove two-dimensional analogs of this notorious conjecture. Once programmed (using my Maple package LADAS) it reproduced their ten theorems, and generated 134 new ones, complete with proofs. All by itself! I believe that the proof of the original 3x+1 conjecture would be in the same vein, but one would need a couple of extra human ideas, and better computers.

Combinatorics,Dynamical Systems

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to make the computer automatically discover and prove mathematical propositions similar to the Collatz 3x + 1 conjecture. Specifically, the author Doron Zeilberger explored how to use the computer - generated method of mathematical research, especially when dealing with problems similar to the Collatz 3x + 1 conjecture, to discover new mathematical laws and proof methods through experimentation and induction.

### Background of the paper
The Collatz 3x + 1 conjecture is a mathematical problem that is simply stated but difficult to prove. Given a positive integer sequence \(x(n)\), its recurrence relation is defined as follows:
\[x(n + 1)=\begin{cases}\frac{x(n)}{2}, & \text{if } x(n) \text{ is even}\\\frac{3x(n)+1}{2}, & \text{if } x(n) \text{ is odd}\end{cases}\]
The initial condition is \(x(0)=x_0\). The conjecture holds that for any positive integer \(x_0\), this sequence will eventually enter the cycle (2, 1).

### Research methods
1. **Experimentation and induction**: The author believes that mathematics is essentially an experimental science, and mathematical laws can be discovered through experimentation. Human mathematicians propose conjectures through experimentation and then verify these conjectures through experimentation.
2. **Computer - aided**: Use the computer to conduct large - scale experiments and generate a large amount of data, thereby discovering potential mathematical laws.
3. **Automated proof**: Develop algorithms and programs to enable the computer to automatically prove these conjectures. For example, use induction to prove certain inequalities or find periodic trajectories through matrix operations.

### Specific steps
1. **Generate conjectures**: Generate a large amount of data through experimentation, observe the laws of the data, and propose conjectures.
2. **Prove conjectures**: Use induction and other mathematical tools to write programs to automatically prove these conjectures.
3. **Verify the proof**: Ensure that the proof generated by the computer is correct.

### Application examples
The author described in detail how to generate and prove problems similar to the Collatz 3x + 1 conjecture through the computer. For example, he considered the second - order recurrence relation of the following form:
\[x(n + 1)=\begin{cases}\frac{\beta x(n - 1)+\alpha x(n)}{2}, & \text{if } x(n - 1)+x(n) \text{ is even}\\\delta x(n - 1)+\gamma x(n), & \text{if } x(n - 1)+x(n) \text{ is odd}\end{cases}\]
where \(\alpha, \beta, \gamma, \delta\in\{- 1,1\}\). Through experimentation and induction, the computer can discover and prove the periodic behavior of these recurrence relations.

### Conclusion
The author demonstrated how to use the computer to generate and prove mathematical conjectures, especially problems similar to the Collatz 3x + 1 conjecture. This method not only improves research efficiency but also may discover new mathematical laws. Although the computer - generated mathematical research is still in its infancy at present, it is expected to make breakthroughs in more complex mathematical problems in the future.