Generative algorithmics is a process of designing and implementing generated algorithms. It belongs in a larger area of algorithm engineering, which deals with experimental evaluation and
speedup of algorithms. At the beginning of the thesis we describe both areas and show the basics of generative algorithmics solving parallel prefix scan. For our main problem we choose
counting k-grams. First, we implement a general algorithm for counting k-grams against which we will compare the generated algorithms. This was followed by the design and implementation of
basic and enhanced generated algorithm. At the end we conducted an experimental evaluation of all the algorithms and verify the effectiveness of generated algorithms. In the last section we
discussed the results, potential improvements and opportunities for further exploration.
|
