Whenever a student reaches the phases, where he must write down his thesis, the question about structure comes up. Kent Beck provides a nice and simple structure (scroll down!). It comes down to "problem, solution, defense, related work".
What is the problem you are solving? Usually something sucks and you can make it better. Provide a good example of suck and also describe it in general. List all requirements any solution must fulfill.
How can you solve the problem? Reuse the examples from the previous section and show how a better solution looks like. Walk the reader through your algorithm in general and with an example. Cover all requirements you listed.
Evaluation sounds better than defense, if you are talking about section titles. While you provided a solution to the problem in the previous section, we still do not know, if your solution is any good. Here you measure, proof, plot, and calculate. Is your solution easy, correct, fast, high-quality, guaranteed, reliable, and more? What makes you so sure about that?
Well done so far, but in science your work must also be novel. Compare your work with others and show what they did not consider. You might be able to reuse your examples once more to show concrete differences.