That's not the thesis itself, just the plan of what I intend working on, what I'm trying to find out, how I intend going about it, and a timeline when I expect certain milestones to be reached. All backed up by evidence that I'll be doing something no-one else had done before, and that it's worth doing, something creative and non-trivial.
Here's just a few of the references needed to back that up. They really should be in a standard format, and will be in my thesis, but this is good enough just for an informal document, a mere proposal:
 IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION, VOL. 1, NO. 1, APRIL 1997 3
Evolutionary Computation: Comments on the History and Current State
Thomas B¨ack, Ulrich Hammel, and Hans-Paul Schwefel
 David B.Fogel quoted at http://www.natural-selection.com/tech_1.html
 Evolution March 2007
EFFECTS OF POPULATION SIZE AND MUTATION RATE ON THE EVOLUTION OF MUTATIONAL ROBUSTNESS
Santiago F. Elena, Claus O. Wilke, Charles Ofria, and Richard E. Lenski
 Natural Computing Vol 1 No 1 2002
Evolution strategies – A comprehensive introduction
Hans-Georg Beyer and Hans-Paul Schwefel
 Antennas and Propagation Society International Symposium, 2000. IEEE , vol.2, no., pp.1034-1037 vol.2, 2000
"Optimum population size and mutation rate for a simple real genetic algorithm that optimizes array factors"
 University of WA Thesis Nov 2007
Optimising Evolutionary Strategies for Problems with Varying Noise Strength
Anthony Di Pietro
 Proceedings of the 2004 Congress on Evolutionary Computation June 2004
Applying evolutionary Algorithms to problems with noisy, time-consuming fitness functions
A.Di Pietro, L.While and L.Barone
 McGill University 2005 Masters Thesis at http://www.music.mcgill.ca/~rebecca/thesis/GAcomparisons/GAcomparisons.htm
Comparing GA population size and mutation rate
 Progress of Theoretical Physics Supplement No.138 (2000) pp. 460-461
Genetic Algorithm Parameter Analysis
 IEEE International Conference on Engineering of Intelligent Systems, April 2006
Genetic Algorithms for Optimal Design of Vehicle Suspensions
Jingjun Zhang; Yanhong Zhang; Ruizhen Gao
 IEEE International Conference on Evolutionary Computation, Sept 2007
Portfolio optimization using multi-objective genetic algorithms
Skolpadungket, P. Dahal, K. Harnpornchai, N.
 IEEE International Conference on Automation and Logistics August 2007
Fast Genetic Algorithms Used for PID Parameter Optimization
Xiangzhong Meng, Baoye Song
 World Congress on Intelligent Control and Automation, June 2006.
Application of Self-Adaptive Genetic Algorithm on Allocating International Demand to Global Production Facilities
Rong-Chang Chen Shiue-Shiun Li Chih-Chiang Lin Tung-Shou Chen
... etc etc
Looking at it.. I've cast my net rather wide, haven't I? Antenna Design, Music, Vehicle Suspension Systems, Portfolio Management, Global Factory Allocation... all could benefit from the work I intend to do. Lots of people seeing different facets of the same basic issue, and finding optimal solutions for particular problems.
I'm good at insights, ideas, concepts. Less good at taking those concepts and putting in the hard yards to see them into fruition, but adequate for that too. Lots of people are, and I'm not amongst the best there. But imaginative concepts, yes, I'm unusually good at coming up with those. Usually ideas of astounding simplicity, so much so that everyone wonders why they didn't think of something as obvious as that before. I'm not bright enough to have complex ideas, I just have good intuition about how to make a complex problem simple.
Evolutionary techniques are great for solving problems where we have no real idea what the answer is, but know a good answer when we see it. The trouble is, we don't know how to go about making good evolutionary computation methods, we take guesses and refine them in particular areas, like portfolio management. In other words, when it comes to making evolutionary algorithms, we have no real idea what the answer is, but know a good answer when we see it. Hmmm... that sounds familiar...
So obviously we should use evolutionary computation to optimise evolutionary computation. That's the first insight. The second insight is on how to map all evolutionary computation methods into a genome, so we can use evolutionary techniques to optimise them. And that's so simple it's trivial.
In fact, it's so simple and obvious, I spent considerable time looking to make sure no-one else had done it before. And as you can see, I looked in a lot of places.
That's the 1% inspiration. Now comes the 99% perspiration. Because now my thesis proposal has been accepted as good enough to get a PhD if written, I better go doing some experimentation and gathering results.
Did I mention that this is fun?