Marwa Ammar

Hybrid harmony search algorithm for global optimization

This paper proposes two hybrid optimization methods based on Harmony Search algorithm (HS) and two different nature-inspired metaheuristic algorithms. In the first contribution, the combination was between the Improved Harmony Search (IHS) and the Particle Swarm Optimization (PSO). The second contribution merged the IHS with the Differential Evolution (DE) operators. The basic idea of hybridization was to ameliorate all the harmony memory vectors by adapting the PSO velocity or the DE operators in order to increase the convergence speed. The new algorithms (IHSPSO and IHSDE) have been compared to the IHS, DE, PSO and some other algorithms like DHS and HSDM. The DHS and HSDM are two existing algorithms, which use different hybridization concepts between HS and DE. All of these algorithms have been evaluated by different test Benchmark functions. The results demonstrated that the hybrid algorithm IHSDE have the better convergence speed into the global optimum than the IHSPSO and the standard IHS, DE and PSO.

Mutational analysis in patients with neuromuscular disorders: Detection of mitochondrial deletion and double mutations in the MT-ATP6 gene.

Mitochondrial diseases encompass a wide variety of pathologies characterized by a dysfunction of the mitochondrial respiratory chain resulting in an energy deficiency. The respiratory chain consists of five multi-protein complexes providing coupling between nutrient oxidation and phosphorylation of ADP to ATP. In the present report, we studied mitochondrial genes of complex I, III, IV and V in 2 Tunisian patients with mitochondrial neuromuscular disorders. In the first patient, we detected the m.8392C>T variation (P136S) in the mitochondrial ATPase6 gene and the m.8527A>G transition at the junction MT-ATP6/MT-ATP8 which change the initiation codon AUG to GUG. The presence of these two variations in such an important gene could probably affect the ATP synthesis in the studied patient. In the second patient, we detected several known variations in addition to a mitochondrial deletion in the major arc of the mtDNA eliminating tRNA and respiratory chain protein genes. This deletion could be responsible of an inefficient translation leading to an inefficient mitochondrial protein synthesis in P2.

Negotiation process for bi-objective multi-agent flexible neural tree model

The major issue of researchers in ANN field is the optimization of the training process including time cost and NN structure. In response to the long training time, Multi-Agent architecture of feed forward Flexible Neural Tree model (MAFNT) is introduced for parallelizing the NN training. Moreover, looking for the best topology of NN, for a given problem, accounts for the large feasible solutions provided. Agents manage different NN structures simultaneously for optimization using Evolutionary Computation algorithms. However, different agents need communications to produce cooperative work and to reach the near-optimum solution. For that, a negotiation process is designed for the multi-agent system. It distributes tasks and organizes the message traffic between agents. They followed negotiation strategy to ensure interactions between themselves, overcoming the difference of NN structures. This model was evaluated through real problem classification datasets. Compared to some existing classifiers, MAFNT shows better performance respecting NN structure complexity and classification rate.

Recurrent Flexible Neural Tree Model for Time Series Prediction

In this paper, a new encoding schemes based on tree representation is represented to encode recurrent multi layer neural network. It implement a learning process formed by two iterative phases: structure optimization and parameters optimization. For the structure evolving, a modified version of the Genetic Programming algorithm was adapted to support the recurrent topology of the network. On the other hand, a hybrid version of Harmony Search algorithm is used to adjust the network parameters including connection weights and neurons parameter set. Besides, the proposed model is evaluated by dynamical chaotic times series and compared with other studies.