A tabu search scheme for abstract problems, with applications to the computation of fixed points (Q2473911)

Defining an abstract problem as a pair \(((X,d),A)\) consisting of a nonempty metric space \((X,d)\) of feasible points and a closed set \(A\subset X\) of acceptable solutions, a tabu search mapping for the abstract problem is any mapping \(N: X\to 2^X\) with open images satisfying (i) \(x\in A\Rightarrow N(X)= \phi\), (ii) \(x\not\in A\Rightarrow x\in N(x)\subset X|A\). The authors propose a general algorithm scheme based on this notion for solving abstract problems in metric spaces in Section~2 and provide a convergence result under a compactness assumption. In Section~3, they specialize their algorithm to the computation of fixed points of continuous mappings from a metric space into itself assuming that the modulus of continuity for the given mapping is available. Section~4 of the paper deals with the particular case of Lipschitzian mappings and Section~5 with that of contraction mappings. It is shown that for contraction mappings, the algorithm scheme includes, as a particular case, the Banach algorithm.