Cuando hablamos de optimización en el ámbito de las ciencias de la computación hacemos referencia al mismo concepto coloquial asociado a esa palabra, la concreción de un objetivo utilizando la menor cantidad de recursos disponibles, o en una visión similar, la obtención del mejor objetivo posible utilizando todos los recursos con lo que se cuenta. Los métodos para encontrar la mejor solución (óptima) varían de acuerdo a la complejidad del problema enfrentado. Para problemas triviales, el cerebro humano posee la capacidad de resolverlos (encontrar la mejor solución) directamente, pero a medida que aumenta la complejidad del problema, se hace necesario contar con herramientas adicionales. En esta dirección, existe una amplia variedad de técnicas para resolver problemas complejos. Dentro de estas técnicas, podemos mencionar las técnicas exactas. Este tipo de algoritmos son capaces de encontrar las soluciones óptimas a un problema dado en una cantidad finita de tiempo. Como contrapartida, requiere que el problema a resolver cumpla con condiciones bastante restrictivas. Existen además un conjunto muy amplio de técnica aproximadas, conocidas como metaheurísticas. Estas técnicas se caracterizan por integrar de diversas maneras procedimientos de mejora local y estrategias de alto nivel para crear un proceso capaz de escapar de óptimos locales y realizar una búsqueda robusta en el espacio de búsqueda del problema. En su evolución, estos métodos han incorporado diferentes estrategias para evitar la convergencia a óptimos locales, especialmente en espacios de búsqueda complejos. Este tipo de procedimientos tienen como principal característica que son aplicables a cualquier tipo de problemas, sin requerir ninguna condición particular a cumplir por los mismos. Estas técnicas no garantizan en ningún caso la obtención de los valores óptimos de los problemas en cuestión, pero se ha demostrado que son capaces de alcanzar muy buenos valores de soluciones en períodos de tiempo cortos. Además, es posible aplicarlas a problemas de diferentes tipos sin mayores modificaciones, mostrando su robustez y su amplio espectro de uso. La mayoría de estas técnicas están inspiradas en procesos biológicos y/o físicos, y tratan de simular el comportamiento propio de estos procesos que favorecen la búsqueda y detección de soluciones mejores en forma iterativa. La más difundida de estas técnicas son los algoritmos genéticos, basados en el mecanismo de evolución natural de las especies. Existen diferentes tipos de problemas, y multitud de taxonomías para clasificar los mismos. En el alcance de este trabajo nos interesa diferenciar los problemas en cuanto a la cantidad de objetivos a optimizar. Con esta consideración en mente, surge una primera clasificación evidente, los problemas mono-objetivo, donde existe solo una función objetivo a optimizar, y los problemas multi-objetivo donde existe más de una función objetivo. En el presente trabajo se estudia la utilización de metaheurísticas evolutivas para la resolución de problemas complejos, con uno y con más de un objetivo. Se efectúa un análisis del estado de situación en la materia, y se proponen nuevas variantes de algoritmos existentes, validando que las mismas mejoran resultados reportados en la literatura. En una primera instancia, se propone una mejora a la versión canónica y mono-objetivo del algoritmo PSO, luego de un estudio detallado del patrón de movimientos de las partículas en el espacio de soluciones. Estas mejoras se proponen en las versiones de PSO para espacios continuos y para espacios binarios. Asimismo, se analiza la implementación de una versión paralela de esta técnica evolutiva. Como segunda contribución, se plantea una nueva versión de un algoritmo PSO multiobjetivo (MOPSO Multi Objective Particle Swarm Optimization) incorporando la posibilidad de variar dinámicamente el tamaño de la población, lo que constituye una contribución innovadora en problemas con mas de una función objetivo. Por último, se utilizan las técnicas representativas del estado del arte en optimización multi-objetivo aplicando estos métodos a la problemática de una empresa de emergencias médicas y atención de consultas domiciliarias. Se logró poner en marcha un proceso de asignación de móviles a prestaciones médicas basado en metaheurísticas, logrando optimizar el proceso de asignación de móviles médicos a prestaciones médicas en la principal compañía de esta industria a nivel nacional. / Tesis doctoral de la Facultad de Informática (UNLP). Grado alcanzado: Doctor en Ciencias Informáticas. Director de tesis: Laura Lanzarini. Co-director de tesis: Guillermo Leguizamón. La tesis, presentada en el año 2013, obtuvo el Premio "Dr. Raúl Gallard" en el 2014.

During domestication, the dog( Canis familiaris), have become skilful in understanding human communication and also in communicating with humans. The wolf ( Canis lupus), is not as skilled with this interspecific communication. When dogs are faced with an unsolvable problem, they seek help from human by e.g. gazing at them. This behaviour has been studied and both age and breed group differences have been showed. In this study, we presented dogs with a task that consisted of a solvable and unsolvable problem in order to see if they gazed at their owner and/or an unfamiliar person for help. Although we did not find any difference in breed groups regarding gazing at humans, we did find that adult dogs (dogs older than 2 years) gazed more frequently at their owner and for a longer duration than adolescent dogs (6 months to 2 years). This may be because the adult dogs have more experience of this communication with humans, as they have lived longer with them. These findings empathize the bond between a dog and its owner that seems to grow stronger during the dogs’ life.

A computer program based upon SPS models developed by D'Zurilla and his colleagues was used (D'Zurilla & Goldfried, 1971; D'Zurilla & Nezu, 1982; D'Zurilla, 1986; D'Zurilla & Nezu, 1990). Comparisons of computerized training, computerized control, workbook training and assessment control groups were conducted on four post-intervention dependent measures: a revised version of the Social Problem-Solving Inventory (D'Zurilla & Nezu, 1990), MeansEnds Problem-Solving Procedure (Platt & Spivack, 1975), SPS SelfEfficacy rating and a Computer/Workbook evaluation. The computerized SPS training group rated their training procedure as significantly more acceptable than those in the workbook training group. Analyses of problem-solving ability and knowledge yielded no significant interactions between problem-solving content and computer contact. No main effects were found for computer contact. A main effect for problem-solving content approached significance. Subjects receiving interventions that contained problem-solving content scored higher on SPS knowledge and ability measures than those who did not.Discussion focuses on reasons for and implications of differences between computerized and non-computerized SPS training groups' performance and preferences. / Department of Psychological Science

This experiment was designed to give further understanding of the underlying factors which influence group idea generation. The first objective of this study was to compare the impact of using computer technology and traditional technologies for creating ideas. The effectiveness of three idea generating techniques, original brainstorming, nominal group technique, and electronic brainstorming were considered. It was, however, hypothesized that electronic brainstorming would outperform the nominal group technique and original brainstorming regardless of the length of time provided.The second objective of this study was to probe how subjects in different idea generating conditions discerned their performance during and after sessions. It was expected that subjects in the original brainstorming groups would perceive that they produce more ideas and they would be more satisfied with results and the process.An ANOVA with a 3x2 factorial design was planned for the study. The independent variables for the study were types of group and types of session. Yet, on account of small sampling size an inferential analysis was precluded. A descriptive analysis was followed.The analysis of five dependent variables, quality, originality, practicality, numbers of nonoverlapping ideas, and perceptions showed that there were no significant differences among three idea generation techniques regarding the length of time provided. However, a quantity variable showed that numbers of nonoverlapping ideas were increased as the length of time were prolonged in six idea generating conditions. / Department of Counseling Psychology and Guidance Services

The purpose of this study was to investigate the effectiveness of two different methods of teaching learning disabled middle school students (6th, 7th, and 8th grades) how to solve one-step addition and subtraction mathematics story problems. This study also compared the generalization of the two instructional methods to problems written in simple syntax which required the performance of two mathematics operations, addition and subtraction, in order to obtain the correct written solution.Teachers were randomly assigned to one of the two instructional methods. The students in the experimental and control classrooms were administered the The Mathematics Computation Screeninq Test, the One-step Story Problem-Solving Test of Mathematics Reasoninq and the Two-step Story Problem-Solving Test of Mathematics Reasoninq. Students who obtained above 80% mastery on the The Mathematics Computation Screening Test and at or below 67% mastery on the pretest of the One-step Story Problem-Solving Test of Mathematics Reasoning were included in the experimental and control groups. Students in the experimental and control groups who meet the above criteria and were at or below the 60% mastery level on the pretest of the Two-step Storv Problem Solving Test of, Mathematics Reasoning were included in the analysis of two-step problems. There were 70 students who meet these criteria. Following administration of the tests, students received 17 days of instruction in one of the two instructional methods.Previous research has shown that good problem-solvers initially have a mental representation of a story problem prior to solving the problem and that accurate performance may be increased by teaching students to generate diagrammatic representations of the problems. Based on this research it was hypothesized that learning disabled students who receive instruction in generating diagrammatic representations would have a higher mean performance on a linear composite of writing number sentences and solving one-step addition and subtraction story problems than learning disabled students who did not receive this instruction when pretest performance on one-step written solutions was held constant. It was also hypothezied that when presented with two-step addition and subtraction story problems learning disabled students who receive instruction in how to generate diagrammatic representations for various one-step addition and subtraction story problems would have a higher mean performance than learning disabled students who do not receive this instruction when pretest one-step and two-step written solutions were held constant.A 2 X 2 X 2 X 2 hierarchical multivariate analysis of covariance mixed effects design followed by examination of step down F ratios was used to test the one-step hypotheses. Analysis of the data indicated no significant difference between the groups on number sentence writing and on solving one-step addition and subtraction story problems varying in syntactic complexity and position of the unknown term. The data did indicate a significant interaction between the within subject factors of syntax, position of the unknown term, and mathematics operation.A 2 X 2 hierarchical analysis of covariance design was used to test the hypotheses regarding generalization of the two instructional methods to two-step story problems of addition and subtraction. Analysis of the data indicated no significant difference between the problem-solving performance of students taught with the diagrammatic instructional method and those taught in the control group. / Department of Special Education

This thesis has examined the hypothesis that sex differences in perception influence anagram problem solving. Specifically, an experiment was designed to study the effects of different letter sizes on the anagram problem-solving performance of male and female eighth grade subjects. It was postulated that males would outperform females under normal letter size conditions, but that females would perform at least as well as males when the anagrams were printed in large letters.The results showed that the female subjects’ performance was superior to the males under both experimental conditions. The results were discussed in terms of lack of experimental control for cognitive style and of possible individual differences in intelligence between the male and female subjects. Ideas were suggested for further research in this area.

One employs the mathematical analysis of decision making when the state of nature is uncertain but further information about it can be obtained by experimentation. Bayesian Decision Theory concerns practical problems of decision making under conditions of uncertainty and also requires the use of statistical and mathematical methods.In this thesis, some basic risk sharing and group decision concepts are provided. Risk is the expected value of the Loss Function of Bayesian Estimators. Group decisions consider situations in which the individuals need to agree both on utilities for consequences and on conditional probability assessments for different experimental outcomes. / Department of Mathematical Sciences

The purpose of this quantitative study was to investigate critical thinking development in an introductory computer programming course in which problem-solving was a key component of the course, compared with another college level computing course in which problem-solving is not a key component. There were two hypotheses in this study. The first was that students would show a greater increase in critical thinking skills after they participate in CINS 113 than CINS 101. The second hypothesis was that students’ critical thinking skills at the beginning of CINS 113 would predict the final grade in the course. Prior to conducting the study, approval was received from both institutions’ review boards and all guidelines were followed. A control group was recruited from students enrolled in a course that was determined to not have problem-solving as a key component of the course and an experimental group was recruited from students enrolled in an introductory computer programming course. Both courses were from a Midwestern community college. Program chairs from various regions throughout the state volunteered their faculty and students to participate in the study. Students were administered the Cornell Critical Thinking Test on the first day of the semester and again at the end of the semester. At the end of the semester, faculty submitted final semester grades for all students participating in the program. A comparison of the pretest was made against the posttest using a repeating ANOVA test to see if there was a significant change between the two scores and if there was a difference in the change in scores between the two groups. In addition, the pretest was analyzed against the final grade for the course to determine if a relationship existed between the critical thinking score at the beginning of the course and the student’s success in the course. A correlational analysis, as well as regression analysis, was conducted. There were a total of 213 students who completed the study. The results of the study supported both hypotheses. / Department of Educational Studies

Thesis (Ph.D. (Clinical Psychology))--North-West University, Potchefstroom Campus, 2006.

The goal of this study is to assess the far-transfer effects of strategy-based working memory (WM) training to a novel problem solving task. Far-transfer refers to the application of trained skills to an untrained situation and is especially important because it deals with the generalization of learning to novel contexts. However, previous working memory training studies have produced little evidence for far-transfer. In the current study, children were trained in two strategies, phonological rehearsal and semantic categorization. These strategies have been suggested to increase the efficiency in processing and encoding of information and are invoked to explain developmental increases in WM capacity. Sixteen 6-to 9-year-olds were randomly assigned to each of four training conditions: semantic and rehearsal training, semantic training only, rehearsal training only, and treated control group. The treated control group performed significantly worse on the problem solving task compared to the three training groups. Surprisingly, the treatment groups did not differ significantly from each other. There was no statistically significant difference in receiving combined training of both strategies compared to only one strategy and furthermore, neither strategy resulted in better performance compared to the other strategy. Future directions for WM training and the implications for cognitive interventions are discussed. / Graduate / 0620 / 0633 / sharonc@uvic.ca