Tabu Search

This is a high level improvement procedure such is utilized to resolve various complex problems. The process starts along with an initial feasible solution and more attempts to determine better solution via investigating among a large neighborhood pool. This also has a special memory termed as "tabu list" such prevents recycling and promotes also the back-tracking of the solution. The generic tabu search algorithm steps of are:

(a)   Generation of initial feasible solution

(b)   Neighbourhood generation

(c)   Generation of Tabu list

(d)   Aspiration criteria set-up

(e)   Stopping criteria

A tabu search procedure's pseudo-code is described in following Figure.

procedure tabu search

begin

select a current point, currentNode, at random

update NodesGenerated bestNode Å currentNode print

repeat

select a new node, newNode, that has the lowest distance in the neighborhood of currentNode that is not on the tabuList, using the twoInterchange method

update NodesGenerated for all nodes checked in neighborhood of currentNode

currentNode <- newNode

if evaluation (currentNode) < evaluation (bestNode)

bestNode <- currentNode

print

until some counter reaches limit

end

                                                              program: Generic Procedure of Tabu Search