This project is concerned with investigating the use of combinators for describing stochastic local search algorithms within pure functional languages, specifically Haskell. The task of effectively combining local search with a pure functional language is one of identifying the key advantages and weaknesses of the functional approach, with relation to local search, and adapting local search to fit within it.
The approach that will be taken here will involve a re-examination of local search in order to identify appropriate atomic characteristics of the algorithms, and appropriate higher order operators to allow their recombination into more complex components.
The current state of the project is that some preliminary research has been done, creating a framework capable of capturing the characteristics of small group of local search algorithms (random walk, iterative improvers, TABU and Simulated Annealing) and this work has been presented in a paper and presentation to IFL2010. An initial library has been created based upon the paper, and can be found in the downloads section.
The current work is looking at a greater variety of local search algorithms and improving the framework, or identifying an alternative, to better support these algorithms.
Unless anyone could not tell, I am not a web designer or web programmer. I hope that this will be a somewhat temporary page, replaced with something more of the form of a customised (but otherwise standard) blog later in the year.page last updated : 2010/09/07