E. Tarantino

Improving search by incorporating evolution principles in parallel Tabu Search

Combinatorial optimization problems require computing efforts which grow at least exponentially with the problem dimension. Therefore, the use of the remarkable power of massively parallel systems constitutes an opportunity to be considered for solving significant applications in reasonable times. In this paper, starting from Tabu Search, a general optimization methodology, a parallel version, oriented to distributed memory multiprocessors and including evolution principles, has been introduced and discussed. The experiments have been performed on classical Traveling Salesman Problems and Quadratic Assignment Problems taken from literature. The results obtained show that the incorporation of evolution principles is very fruitful for the search strategy in terms of both convergence speed and solution precision.<<ETX>>

SIMULATION OF GENETIC ALGORITHMS ON MIMD MULTICOMPUTERS

In this paper, a Parallel Genetic Algorithm has been developed and mapped onto a coarse grain MIMD multicomputer whose processors have been configured in a fully connected chordal ring topology. In this way, parallel diffusion processes of good local information among processors have been carried out. The Parallel Genetic Algorithm has been applied, specifically, to the Travelling Salesman Problem. Many experiments have been performed with different combinations of genetic operators; the test results suggest that PMX crossover can be avoided by using only the inversion genetic operator and that a diffusion process leads to improved search in Parallel Genetic Algorithms.

Artificial Neural Networks Optimization by means of Evolutionary Algorithms

In this paper Evolutionary Algorithms are investigated in the field of Artificial Neural Networks. In particular, the Breeder Genetic Algorithms are compared against Genetic Algorithms in facing contemporaneously the optimization of (i) the design of a neural network architecture and (ii) the choice of the best learning method for nonlinear system identification. The performance of the Breeder Genetic Algorithms is further improved by a fuzzy recombination operator. The experimental results for the two mentioned evolutionary optimization methods are presented and discussed.

An analysis of parallel heuristics for task allocation in multicomputers

In literature there exist many heuristic optimisation techniques which have been proposed as general-purpose methods for solving difficult problems. Of course, the question which of them is more powerful is in general meaningless, however, their application and comparison on real, well-limited problems is quite interesting and intriguing. Furthermore, parallel versions for such techniques are welcome, allowing to reduce the search times or to find new innovative solutions unreachable in a sequential environment. Within this paper we describe two such techniques, the Genetic Algorithms and the Simulated Annealing, and provide a general parallelisation framework for heuristic methods which is based on a locally linked search strategy. A comparative analysis of the parallel versions of these techniques is performed on the solution of a set of different-sized Task Allocation Problems in terms of better absolute solution quality, of lower convergence time to a same solution and of robustness expressed as lower variance around the mean value.

Testing parallel evolution strategies on the quadratic assignment problem

Parallel evolution strategies are demonstrating to be worthwhile in a variety of contexts. In this paper, besides the classical genetic and evolutionary strategies, a hybrid evolutionary approach which incorporates memory of the search history within the structure is analyzed. The parallel evolution algorithms are mapped on a distributed memory MIMD multicomputer whose processors are configured in a torus topology. The simulations are conducted using the quadratic assignment problem as an artificial environment. The relationship between genetic representations and recombination operators is investigated. The experimental results obtained show the value of structures richer than bit strings and the effectiveness of memory for the evolution process.<<ETX>>

MijnMutation Operator for Aerofoil Design Optimisation

A new mutation operator, called M ijn , capable of operating on a set of adjacent bits in one single step, is introduced. Its features are examined and compared against those of the classical bit-flip mutation. A simple Evolutionary Algorithm, M-EA, is described which is based only on selection and M ijn This algorithm is used for the solution of an industrial problem, the Inverse Aerofoil Design optimisation, characterised by high search time to achieve satisfying solutions, and its performance is compared against that offered by a classical binary Genetic Algorithm. The experiments show for our algorithm a noticeable reduction in the time needed to reach a solution of acceptable quality, thus they prove the effectiveness of the proposed operator and its superiority to GAs for the problem at hand.

Parallel tabu search versus parallel evolution strategies

There exists in scientific, industrial and financial communities a very strong request for techniques able to efficiently solve complex optimization problems. Because of this, several techniques are being currently investigated. Among them evolutionary algorithms and tabu search seem very interesting, not only for their intrinsic features but also because they both are easily parallelizable, so that they can take advantage of the parallel machines available on the market. A new parallel approach to tabu search (PTS) is introduced and compared against parallel evolution strategies on classical optimization problems taken from literature. The experimental results have shown the superiority of the PTS in both the solution quality and the convergence time.<<ETX>>

The effectiveness of co-mutation in evolutionary algorithms: the /spl Mscr//sub ijn/ operator

A new mutation operator, designated /spl Mscr//sub ijn/, which is capable of operating on a set of adjacent bits in one single step, is introduced. Its properties are examined and compared against those of the bit-flip mutation. A simple evolutionary algorithm is described which is based only on selection and /spl Mscr//sub ijn/. This algorithm is used for the optimization of a well-known problem testbed, and its performance is compared against that offered by both a classical genetic algorithm and a more sophisticated one. The obtained results prove the effectiveness of the /spl Mscr//sub ijn/ operator.

Comparing Parallel Tabu Search and Parallel Genetic Algorithms on the Task Allocation Problem

There exist many sequential heuristic combinatorial techniques to efficiently solve a wide set of problems. Researchers are proposing new parallel versions for them so as to take advantage of the power offered by parallel computers. In this paper new locally-linked parallel versions for two such techniques, Tabu Search and Genetic Algorithms, are proposed and tested on different-sized items of the Task Allocation Problem, one of the most challenging optimisation problems.

Mapping Parallel Genetic Algorithms on WK-Recursive Topologies

In this paper a parallel simulator of Genetic Algorithms is described. The target machine is a parallel distributed-memory system whose processors have been configured in a WK-Recursive topology. A diffusion mechanism of useful local information among processors has been carried out. Specifically, simulations of genetic processes have been conducted using the Travelling Salesman Problem as an artificial environment. The experimental results are presented and discussed. Furthermore, performance with respect to well-known problems taken from literature is shown.