An algebraic approach to mechanical tabu search algorithm generation

Traditional approaches to apply tabu search method typically require formulating an algorithmic structure for each individual problem. Based on algebraic specifications, the paper presents a unified and mechanical framework for implementing tabu search algorithms for combinatorial optimization problems. We define a generalized specification using high-order functions to describe search strategies, from which efficient algorithms can be generated for concrete problems by data type refinement and functional refinement. We also extend the framework for multi-objective problems by enriching specification behaviors. Our approach achieves a high level of abstraction and mechanization without losing performance.