Meta-heuristic search algorithms, by their very nature, are applicable across a range of optimisation problems. In practice however, meta-heuristics have been tailored to solve particular problems. Recent work by Randall and Abramson (2001b)has successfully shown that iterative meta-heuristics, such as simulated annealing and tabu search, can be successfully generalised to solve a range of problems without modification though the use of a uniform representation language. Constructive meta-heuristics, such as ant colony optimisation and generalised random adaptive search procedures, pose more substantial problems to achieve this same level of generalistaion. This paper investigates the issues involved and suggests some measures by which generalisation could be achieved.
|Title of host publication||Operations Research/Management Science at Work|
|Editors||Erhan Kozan, Azuma Ohuchi|
|Place of Publication||Boston|
|Publisher||KLUWER ACADEMIC PUBL|
|Number of pages||18|
|Publication status||Published - 2002|
Randall, M. (2002). A General Framework for Constructive Meta-heuristics. In E. Kozan, & A. Ohuchi (Eds.), Operations Research/Management Science at Work (pp. 111-128). Boston: KLUWER ACADEMIC PUBL. https://doi.org/10.1007%2F978-1-4615-0819-9_7