In this paper we present a classification of parallel tabu search metaheuristics based, on the one hand, on the control and communication strategies used in the design of the parallel tabu search procedures and, on the other hand, on how the search space is partitionned. These criteria are then used to review the parallel tabu search implementations described in the literature. The taxonomy is further illustrated by the results of several parallelization implementations of a tabu search procedure for multicommodity location-allocation problems with balancing requirements. Key words: Tabu search metaheuristics, Parallelization strategies, Taxonomy R'esum'e Nous pr'esentons un sch'ema de classification des algorithmes parall`eles de recherche avec tabous. La taxonomie est bas'ee, d'une part, sur les strat'egies de controle et de communication des algorithmes parall`eles de recherche avec tabous et, d'autre part, sur les r`egles de partitionnement du domaine. Ces crit`eres sont ensuite...

We study and compare asynchronous parallelization strategies for tabu search, and evaluate the impact on performance and solution quality of some important algorithmic design parameters: number of processors, handling of exchanged information, etc. Parallelization approaches are implemented and compared by using a tabu search algorithm for multicommodity location-allocation problems with balancing requirements. Key words: Tabu search methods, Parallel algorithms, Asynchronous strategies, Multicommodity location-allocation with balancing requirements R'esum'e Nous 'etudions et comparons plusieurs m'ethodes parall`eles asynchrones de recherche avec tabous. Nous identifions les approches de parall'elisation les plus prometteuses et nous 'evaluons l'impact sur le comportement des m'ethodes et sur la qualit'e des solutions de plusieurs aspects importants du design algorithmique: nombre de processeurs, traitement de l'information 'echang'ee, etc. Les approches de parall'elisation sont exa...

Tabu search has achieved widespread successes in solving practical optimization problems. Applications are rapidly growing in areas such as resource management, process design, logistics, technology planning, and general combinatorial optimization. Hybrids with other procedures, both heuristic and algorithmic, have also produced productive results. We examine some of the principal features of tabu search that are most responsible for its successes, and that offer a basis for improved solution methods in the future. Note: This expanded version contains additional illustrations and information on candidate list strategies, probabilistic tabu search, strategic oscillation and parallel processing options. Sections have also been added on principles of intelligent search.

We study and compare synchronous parallelization strategies for tabu search. We identify the most promising parallelization approaches, and evaluate the impact on performance and solution quality of some important algorithmic design parameters: length of the synchronization steps, number of processors, handling of exchanged information, etc. Parallelization approaches are implemented and compared by using a tabu search algorithm for multicommodity location-allocation problems with balancing requirements. Key words: Tabu search methods, Parallel algorithms, Synchronous strategies, Multicommodity location-allocation with balancing requirements R'esum'e Nous 'etudions et comparons plusieurs m'ethodes parall`eles synchrones de recherche avec tabous. Nous identifions les approches de parall'elisation les plus prometteuses et nous 'evaluons l'impact sur le comportement des m'ethodes et sur la qualit'e des solutions de plusieurs aspects importants du design algorithmique: fr'equence de syn...

The ability of neural networks to closely approximate unknown functions to any degree of desired accuracy has generated considerable demand for Neural Network research in Business. The attractiveness of neural network research stems from researchers' need to approximate models within the business environment without having a priori knowledge about the true underlying function. Gradient techniques, such as backpropagation, are currently the most widely used methods for neural network optimization. Since these techniques search for local solutions, a global search algorithm is warranted. In this paper we examine a recently popularized optimization technique, Tabu Search, as a possible alternative to the problematic backpropagation. A Monte Carlo study was conducted to test the appropriateness of this global search technique for optimizing neural networks. Holding the neural network architecture constant, 530 independent runs were conducted for each of seven test functions, including a pr...

Roughly speaking, parallel local search techniques can be divided into three categories: low-level parallelization strategies (e.g., master-slave schemes), solutionspace partitioning methods and multi-thread procedures in which several processes explore concurrently the same search space. The multi-thread technique can be further subdivided into independent and cooperative search thread algorithms. In this paper, we focus on cooperative multi-thread heuristics applied to methods such as tabu search, and attempt to identify the key questions to be addressed in the design of any algorithm in this class. In particular, we show that questions related to inter-agent communications are a central element of the algorithmic design of these methods. 1 Introduction One of the simplest and most often used techniques to parallelize iterative heuristic search methods such as tabu search (TS) performs many concurrent independent search threads on the same problem. This approach have been ...

Distributed computer systems may be used in various ways to nd good solutions to dicult combinatorial optimization problems. An interesting approach consists in executing concurrently dierent search methods that exchange information gathered in previously explored regions of the solution space. This cooperation mechanism strongly impacts how the solution space is explored. We introduce a formal representation of such cooperative procedures based on discrete-time dynamical systems. We describe how the search behavior of cooperative programs depends on systems of complex and correlated interactions. We derive the notion of systemic cooperation and illustrate through simulations the relevance of this notion to the understanding of the search behavior of cooperative procedures.

We present the main freight transportation planning and management issues, briefly review the associated literature, describe a number of major developments, and identify trends and challenges. In order to keep the length of the paper within reasonable limits, we focus on long-haul, intercity, freight transportation. Optimization-based operations research methodologies are privileged. The paper starts with an overview of freight transportation systems and planning issues and continues with models which attempt to analyze multimodal, multicommodity transportation systems at the regional, national or global level. We then review location and network design formulations which are often associated with the long-term evolution of transportation systems and also appear prominently when service design issues are considered as described later on. Operational models and methods, particularly those aimed at the allocation and repositioning of resources such as empty vehicles, are then described....

Abstract- The search history of memory based heuristics like tabu search can be used to design a category of parallel algorithms called cooperative search. These algorithms execute in parallel several search programs on the same op-timization problem instance. At run time, the data gathered in the memory by one sequen-tial search program need not be used only by this program, but it can be recycled and shared with other concurrently executing tabu search programs for the same purpose. Several such cooperative search algorithms have already been proposed in the literature, but yet, very little is known about how they work as problem solving methods, particularly about their optimization properties. In this paper we compare the global behavior of cooperating and non-cooperating tabu search programs. We show that cooperating programs tend to have a search pattern whichis less diversified than non-cooperating programs. Our findings also indicate that this second or-der impact of the sharing of gathered data on the search behaviors of cooperating programs is not related to the optimization properties of the individual tabu programs. 0-7803-4778-1 /98 $10.00 0 1998 IEEE 2379 I.

In this paper, a tabu search heuristic is combined with slope scaling to solve a discrete depot location problem, known as the multicommodity location problem with balancing requirements. Although the uncapacitated version of this problem has already been addressed in the literature, this is not the case for the more challenging capacitated version, where each depot has a fixed and finite capacity. The slope scaling approach is used during the initialization phase to provide the tabu search with good starting solutions. Numerical results are reported on various types of large-scale randomly generated instances. The quality of the heuristic is assessed by comparing the solutions obtained with those of a commercial mixed-integer programming code.