Search

This paper presents a new methodology based on the application of Support Vector Machine algorithms, Naïve Bayes and Genetic Algorithms in diagnostics of psychosocial evaluations for the identification and prediction of the psychosocial risk level of public-school teachers in Colombia. A comparative study of the model of machine learning for prediction was carried out: Support Vector Machines (SVM) and Naïve Bayes, in two stages, first with all the variables and second, reducing the dimensionality of the database applying genetic algorithms, The best forty variables with the best efficiency in prediction accuracy were selected. The database used consisted of 3000 epidemiological records, which corresponded to teachers from public schools in the metropolitan area of a Colombian city. The use of SVM easily detected variables of physiological type and the best prediction performance was obtained with accuracy of 96.3%. [ABSTRACT FROM AUTHOR]

This paper presents a new methodology based on machine learning techniques in diagnostics of psychosocial assessments to identify the risk level in teachers of public schools in Colombia. A comparative study of three important models of machine learning for prediction was done: artificial neural networks, decision trees and naive bayes, reducing the dimensionality of the data. This was done by applying genetic algorithms, algorithm of the expected amount of information, the algorithm GainRatioAttributeEval, Pearson's relation coefficient and principal components analysis. A database was used with 5340 epidemiological records, corresponding to psychosocial evaluations of teachers from public schools in the metropolitan area of a Colombian city. The best predictive performance was obtained with the model of artificial neural networks with an accuracy 93%. [ABSTRACT FROM AUTHOR]

This paper presents a comparison of metaheuristics applied to the power system interdiction problem. This problem considers the interaction of a disruptive agent and a system operator. The disruptive agent aims at maximizing the damage of the power system, measured as load shedding. This agent has limited resources and must select the combination of elements to attack that result in the greatest damage to the system, anticipating the reaction of the system operator. On the other hand, the system operator must minimize the load shedding by redispatching generation resources. The interdiction model is nonlinear, nonconvex and multimodal. The solutions provided by three metaheuristic techniques are compared: Genetic Algorithm, GRASP and Iterated Local Search. Results show that the Iterated Local Search adapts better to the type of problem and allows obtaining the best rate between quality of solutions and computation time. [ABSTRACT FROM AUTHOR]

This paper proposes a methodology based on decentralized operation and coordination of large power systems, obtaining the system's optimization in a simplified way, and reducing time consumption. The optimal search of generator voltage, tap changer and compensation, is accomplished by using an evolutionary technique, and considering the problem as a multi-objective optimization case. In this paper, both the decomposition network and the optimization problem are applied to two study cases (39 and 190 buses). Global optimal operation is attained, preserving the voltage stability and the exchanges among areas, with minimal operational information, simplicity in the equivalents and time consuming reduction. Results are similar to those obtained by a centralized optimization strategy. [ABSTRACT FROM AUTHOR]

In this paper an innovative focus to programming Job Shop systems, based on sustainability, is presented. The following sustainability variables are evaluated: carbon dioxide emissions (environmental variable), sound intensity, accident rate and handled weight (social variables: occupational health) and finally, makespan time and direct labor costs (economic variables). The methodology is structured in fourteen steps that cover aspects such as multiobjective method design, description of genetic algorithms, and comparison with other methods. It is determined that the designed method shows greater performance for the variable weighted factor P(Xj) when they are compared with results from other methods such as programming rules and mobile bottle neck method. [ABSTRACT FROM AUTHOR]

This paper presents an assessment of optimal location and sizing of distributed generation in distribution systems. For such an assessment the following search techniques were implemented: i) simulated annealing, ii) variable neighborhood search, iii) genetic algorithm and iv) hybrid genetic algorithm. Two objectives have been considered in the optimal location and sizing problem: the minimization of electric power losses and the improvement of voltage profile. To compare the effectiveness of the methodologies a 34 bus distribution test system is used. The correct location and sizing of the distributed generation allowed substantially improving the voltage profile of the network and reducing in 80% the power losses in the test system. [ABSTRACT FROM AUTHOR]

This paper presents a methodology for optimal Distributed Generation (DG) allocation using a hybrid genetic algorithm. In this case, the objective function is the minimization of power losses. The proposed algorithm involves an artificial neural network to evaluate the fitness function and a local search subroutine that allows the algorithm to explore a wider search space, and eventually, escape from local optimal solutions. The proposed approach was tested on a 34-bus system showing that the proper allocation of DG units allows the reduction of active power losses and the improvement of voltage profile. [ABSTRACT FROM AUTHOR]

In this paper compares studies over routing and wavelength assignment for better some services indicators, such as blocking probability and network utilization. In modern telecommunications are gradually subjected to increased demand, which results in higher requirements in photonic transport networks. This phenomenon forces us improve routing systems, in order to better serve current demand and have capacity of attention of future demand. The comparison is made between the algorithms Genetic Algorithms and Simulated Annealing, both simulated in the optical network NSFNET using wavelength division multiplexing WDM. The results showed that heuristic algorithms are better in dynamic demand and high charge, functioning better under stress. It is concluded that genetic algorithms provide the best solution for future demand. [ABSTRACT FROM AUTHOR]

In this paper, the performance of the variables: makespan time, defective fraction due to operator fatigue and direct labor cost, in a job shop manufacturing system is presented. Two methods are compared; the first one blends elements from VEGA and MOGA algorithms (V&M), while the second one corresponds to a weighted sums method (WSM). Applying these methodologies, it was found that the method V&M has a better performance than the WSM. Comparing he variable makespan time (in hours) the first is better by 7.0 %; comparing the variable makespan time (in weeks) by 6.7 %; and comparing the variable defective fraction by 1.4 %. Regarding the variable labor costs, both methods have the same performance. [ABSTRACT FROM AUTHOR]

In this paper the finite element method is applied to the analysis of a three-phase magnetic induction motor squirrel-cage. Different geometries have been applied in the design of the rotor. These geometric variations of the rotor create an electromagnetic profile for each type of geometry. To study the rotor motion, a frequency analysis of the model has been done. The identifying parameters deduced from the finite element method, are adjusted by evolutionary genetic algorithms. Thus, the parameters in the circuit model were identified for each geometry under consideration. A check of the torque was done to validate the model, verifying the accuracy of the method. [ABSTRACT FROM AUTHOR]

This paper presents the application of an evolutionary algorithm using Matlab to obtain the optimum main angles of the cutting geometry of a machining tool. The optimization method considered the non linear model of Kaldor and Venuvinod. Different algorithm configurations were tested using data obtained from the literature. Several simulations were carried out to validate the performance of the algorithm and to show the usefulness of the method employed. It is concluded that that the proposed approach shows to be stable and coherent in the results for the cases studied. [ABSTRACT FROM AUTHOR]