Typical technical writing plan:

0.1 Title that reflects essence of problem & algorithm
0.1 Abstract that describes problem and novel approach & results
1.1 Introduction to problem 
1.2 Problem definition 
1.3 Why algorithm is original & novel
2.1 Technical description of algorithm
2.2 Computational mathematical & statistical foundations.
2.3 Pseudo-code detailing algorithm
3.1 Description of simple, motivating, concrete example
3.2 Step-through algorithm on example
4.1 Description of implementation, availability of code.
4.2 Details of algorithms used, language etc.
5.1 Describe with/without testing criterion
5.2 Provide statistics and details of testing examples
5.3 Display comparative results graphically
5.4 Explain and interpret results

0 Title  & abstract:

Concisely describes key idea and why approach is unique,
often may involve the integration of two things.

1. Problem definition:

Requires problem definition,
e.g. simply describing techniques in reasoning etc. may not contribute anything new or be novel

2. Technical details of algorithm:

Keep in mind that all too often, 
technical writing is too shallow, needs more motivation.
lacks technical detail and needs more technical content

3. Concrete examples:

Good, concrete examples are ideally required, without requiring reader to think about too many levels of abstraction (warning: good simple, illustrative examples are often the hardest examples to come up with!)

4. Implementation details:

Technical writing generally requires at least some implementation details.

5. Evaluation of results of comparative experiments:

Generally, you are expected to show results of experiments and comment
and importantly interpret the results.
All too often, evaluation is missing. Good writing includes comments on the validity and efficiency of approaches.
Comparative results are a good way to proceed experimentally, you can do 
 a combination of experiments (with, without, etc.).