• United States
by Nicolas Logan

Computing naturally: Benefits and risks of genetic algorithms

Apr 18, 20114 mins

Nicholas K. Logan, CEH is a graduating senior in the information assurance program of the School of Business and Management at Norwich University. As one of his essays in the IS342 Management of Information Assurance course, he wrote about potential benefits and risks of  several types of computing based on biological models. Everything that follows is Mr. Logan’s work with minor edits.

* * *

The use of naturally occurring systems for massively parallel computing could change the nature of encryption methods, allow computers to connect with living tissue, and allow for computational systems that evolve independently of their creators’ intent. The risks are greater than any other posed by computer systems up to now, but the potential benefits are great. These new systems are in development now in labs around the world. The day is coming where computers can be grown, not built.

Genetic Algorithms

The current world of biologic computation can be divided into three categories: genetic algorithms, natural computation, and nanotechnology.

Genetic algorithms are very different from the others since they do not utilize natural systems directly. Genetic algorithms are functions that utilize processes observed in nature to produce algorithms and solutions that survive the evolutionary constrains in which the algorithm was produced. Genetic algorithms start with random combinations of basic functions and through placing evolutionary goals on the produced algorithms it will choose and alter or recombine those algorithms that weren’t as wrong as the others and run this new generation of algorithms. Given enough time and computer power genetic algorithms can solve many issues that took centuries to solve and find solutions to problems outside of current human comprehension.

The use of genetic algorithms has recently shown its strength by allowing a computer to compute, “the law of conservation of momentum, and Newton’s second law of motion,” with only input about pendulums and basic mathematical functions (e.g., addition, subtraction). 

Risks of Genetic Algorithms

Genetic algorithms on their own are not necessarily inherently risky, but when used with the technology in the following two articles, DNA-based computation and nanotechnology, there is a possibility for technology to leave humanity’s control and take control of its own destiny.

[MK adds: Geek cartoonist Randall Munroe has a cute cartoon about the possible perils of genetic algorithms in which he advises programmers to include the possible costs of “thisAlgorithmBecomingSkynetCost” (a reference to the Terminator films) in which rogue computers called Skynet try to wipe out humanity.]

If robots with the ability to alter biologic material and to carry out massively parallel computation were able to provide parameters to improve themselves using genetic algorithms, the growth rate in capabilities could increase at an exponential rate. One form of exponential self improvement could go as follows:

1. Genetic algorithm used to improve hardware;

2. Improved hardware allows software to run faster;

3. Faster software develops better version of the software;

4. Recursively follow steps 1 through 3 until resources run low;

5. Use genetic algorithm to solve current issue of resource scarcity;

6. Go to step one.

In the next of the three articles in this series, Mr. Logan looks at DNA-based computation.

* * *

Nicholas K. Logan, CEH, is a member of the Norwich University Corps of Cadets. After he graduates with his BSc in Computer Security and Information Assurance in May 2011, he will be working for a large Washington, D.C. area consulting firm where he has been an intern working on risk management Monte Carlo modeling. He is a member of the Association for Computing Machinery and has been inducted into the Upsilon Pi Epsilon honor society. In addition to his wide interests in information security and risk management, he is fascinated by computational complexity theory, artificial intelligence, and evolutionary theory.


Please support Norwich University ROTC student Zach Wetzel’s fund-raising run for the Semper Fi Injured Marines Fund.