The goal of this paper is to discuss the tabu search (TS) meta-heuristic and its enhancement for combinatorial optimization problems. Firstly, the issues related to the principles and specific features of the standard TS are concerned. Further, a promising extension to the classical tabu search scheme is introduced. The most important component of this extension is a special kind of diversification mechanism. We give the paradigm of this new improved TS strategy, which is called an iterated tabu search (ITS). ITS was applied to the difficult combinatorial optimization problems, the traveling salesman problem (TSP) and the quadratic assignment problem (QAP). The results of the experiments with the TSP and QAP show the high efficiency of the ITS strategy. The outstanding performance of ITS is also demonstrated by the fact that the new record-breaking solutions were found for the hard QAP instances - tai80a and tai100a.

No abstract

Today exchange of data among CAD systems becomes more important. As each CAD system has its own flavor to represent the same objects, exchange of data among them is full of issues. Most common problems are incomplete transfer of presentation information, associativity between representation and presentation, usage of annotation planes and text. They are analyzed in details and possible solutions are suggested as well.

In this paper, we investigate some modified local search (LS) heuristics for the solution of symmetric traveling salesman problem (TSP). These modifications are mainly due to the use of extended neighborhood structures. In addition, we are concerned with several new sets of the moves (transitions of solutions) based on the extended configurations of edge exchanges. We are also examining the performance of these extensions being used in an iterated local search (ILS) paradigm. The results from the experiments with the benchmark TSP instances from the TSP library (TSPLIB) demonstrate that the introduced improvements enable to seek solutions of higher quality without substantially increasing computational complexity.