Omar Abdelkafi

Using a novel parallel genetic hybrid algorithm to generate and determine new zeolite frameworks

Abstract Zeolite structure determination and zeolite framework generation are not new problems but due to the increasing computer power, these problems came back and they are still a challenge despite the recent progress in terms of structural resolution from X-rays and electron diffraction. The infinite number of potential solutions and the computational cost of this problem make the use of metaheuristics significant for this problem. In this paper, we propose a new approach based on parallel genetic hybrid algorithm for zeolites using a modified modelization of the objective function to find hypothetical zeolite structures, close to the thermodynamic feasibility criterion. A population made of random atoms is initialized. At each generation, a crossover operator and a mutation heuristic are applied. Each individual of the population generates a potential zeolitic structure by applying the symmetry operators of a given crystallographic space group. This structure is evaluated with our objective function. From the unit cell parameters and the number of T atoms in the asymmetric unit, 6 possible zeolitic interesting structures have been found.

Comparison of Two Diversification Methods to Solve the Quadratic Assignment Problem

The quadratic assignment problem is one of the most studied NP-hard problems. It is known for its complexity which makes it a good candidate for the parallel design. In this paper, we propose and analyze two parallel cooperative algorithms based on hybrid iterative tabu search. The only difference between the two approaches is the diversification methods. Through 15 of the hardest well-known instances from QAPLIB benchmark, our algorithms produce competitive results.

A Survey on the Metaheuristics Applied to QAP for the Graphics Processing Units

The computational power requirements of real-world optimization problems begin to exceed the general performance of the Central Processing Unit (CPU). The modeling of such problems is in constant evolution and requires more computational power. Solving them is expensive in computation time and even metaheuristics, well known for their eficiency, begin to be unsuitable for the increasing amount of data. Recently, thanks to the advent of languages such as CUDA, the development of parallel metaheuristics on Graphic Processing Unit (GPU) platform to solve combinatorial problems such as the Quadratic Assignment Problem (QAP) has received a growing interest. It is one of the most studied NP-hard problems and it is known for its high computational cost. In this paper, we survey several of the most important metaheuristics approaches for the QAP and we focus our survey on parallel metaheuristics using the GPU.

Distributed Multistart Hybrid Iterative Tabu Search

The quadratic assignment problem (QAP) is one of the most studied NP-hard problems. It is a problem known for its computational cost which makes it a good candidate for parallel and distributed design. In this paper, we propose a new Distributed Multistart Hybrid Iterative Tabu Search (DMHITS). This algorithm follows the design of the algorithmic level. Through 34 of the hardest well-known instances from QAPLIB benchmark, the DM-HITS can get the best known solution for almost all the instances. From the 340 runs on these benchmark instances, our algorithm gets more than 300 times the best known solution. This experimentation shows that our proposed algorithm can exceed or equal six leading algorithms from the literature.

MEmory Genetic Algorithm Hybridized for Zeolites

Zeolite structure determination is an interesting challenge even with the progress in terms of structural resolution from X-rays and electron diffraction. The infinite number of potential solutions and the computational cost of this problem make the use of an evolutionary algorithm significant for this challenge. In this paper, we propose a new parallel and distributed hybrid genetic algorithm called MEmory Genetic Algorithm Hybridized for Zeolite (MEGA-HZ). This experimentation shows that the proposed algorithm is able to satisfy the constraints of the objective function to determine viable zeolite structures. From the 6 unit cell parameters and density, the MEGA-HZ has found 6 different viable zeolite structures.

Improved Hybrid Iterative Tabu Search for QAP Using Distance Cooperation

The quadratic assignment problem can be considered as one of the hardest and most studied combinatorial problems. In this paper, we propose and analyze three distributed algorithms based on hybrid iterative tabu search. These algorithms follow the design of the parallel algorithmic level. A new mechanism to exchange information between processes is introduced. Through 34 well-known instances from QAPLIB benchmark, our algorithms produce competitive results. This experimentation shows that our best propositions can exceed or equal several leading algorithms from the literature in almost all the hardest benchmark instances.

Fast Hybrid BSA-DE-SA Algorithm on GPU

This paper introduces a hybridization of Backtracking Search Optimization Algorithm (BSA) with Differential Evolution (DE) and Simulated Annealing (SA) in order to improve the convergence speed of BSA. An experimental study, conducted on 20 benchmark problems, shows that this approach outperforms BSA and two other hybridizations [4, 18], in terms of solution quality and convergence speed. We also describe our CUDA implementation of this algorithm for graphics processing unit (GPU). Experimental results are reported for 10 high-dimensional benchmark problems, and it highlights that significant speedup can be achieved.

Data Exchange Topologies for the DISCO-HITS Algorithm to Solve the QAP

Exchanging information between processes in a distributed environment can be a powerful mechanism to improve results for combinatorial problem. In this study, we propose three exchange topologies for the distance cooperation hybrid iterative tabu search algorithm called DISCO-HITS. These topologies are experimented on the quadratic assignment problem. A comparison between the three topologies is performed using 21 well known instances of size between 40 and 150. Our algorithm produces competitive results and can outperform algorithms from the literature for many benchmark instances.

Multi-level Parallelization for Hybrid ACO

The Graphics-Processing-Unit (GPU) became one of the main platforms to design massively parallel metaheuristics. This advance is due to the highly parallel architecture of GPU and especially thanks to the publication of languages like CUDA. In this paper, we deal with a multi-level parallel hybrid Ant System (AS) to solve the Travelling Salesman Problem (TSP). This multi-level is represented by two parallel platforms. The first one is the GPU, this platform is used for the parallelization of tasks, data, solution and neighborhood-structure. The second platform is the MPI which is dedicated to the parallelization of programs. Our contribution is to use these two platforms to design a hybrid AS with a Local Search and a new heuristic.