José Guadix Martín

Mixed Trips in the School Bus Routing Problem with Public Subsidy

In some regions, school buses have a subsidy associated with the number of school buses needed in a day, in relation to the number of schools or students. This paper presents the application of a Tabu Search (TS) metaheuristic to generate an information system that will allow searching for the best selection of routes and schedules with the minimum number of buses. The complexity and magnitude of the problem is important, especially at economic level for Andalusia government, because there are many buses and roads to go to school. The approach is more relevant for the necessary austerity in public services, considering that the quality is guaranteed by a set of restrictions. In the mixed trip problem, students of different schools can share the same bus at the same time. We use a mixed load algorithm for the school bus routing problem (SBRP) and measure its effects on the number of required vehicles. The solution could reduce the required number of vehicles compared with the current practice.

Neural Network Application for Risk Factors Estimation in Manufacturing Accidents

In occupational safety, when a neural network is trained, it is possible to predict the outcome given a combination of risk factors. Risk assessment is probably the most important issue in occupational safety. Risk assessment facilitates the design and prioritization of effective prevention measures. Neural network were applied for predicting the severity of accidents, which is important to assess risks. Data sets were obtained from the official accident notifications in the manufacturing sector of Andalusia in 2011. The results confirm that neural networks are useful in risk factor estimation. Association analysis was used to identify the most important risk factors within the predicting variables. Diagnostic array analyses show that for preventive purposes it is better to use a reduced data set with a case-control approach in order to improve the specificity and the sensitivity.

Short-Term Forecasting in Office Consumption with Identification of Patterns by Clustering

Fuzzy clustering algorithm has demonstrated advantages in data mining, especially in problems with large collections of imprecise or fuzzy data. This methodology is widely used in the context of pattern recognition. It is proposed this methodology to use in the forecast of the comfort area of building or in the forecasting of consumption. In this paper the concept of fuzzy clustering, which is widely used in the context of pattern recognition. Based on the study of the fuzzy algorithm, we propose a method to forecast the comfort and consumption.