I worked with logic fuzzy, ANFIS, Neuronal network and particle swarm, all of this techniques to analyze and simulate berry borer behaviour in colombian crops.
In this paper a simulation of a basic model of multiple swarms of particles is presented, illustrating the behavior of the coffee berry borer infestation in Colombian coffee crops generating information. In regard of this problem it is used engineering techniques, which aims to promote impact of the level at which the pest harms the national economy. The results show an adequate performance of the model and simulation proposed.
In this paper the performance and time simulation of a model of multiple swarms of particles is analyzed, comparing their results on a physical machine versus a virtual machine. For experiments they were executed several cases taking the time for each simulation (each scenario was executed 50 times). The implementation of this virtualization allowed significantly reduce the execution time of the experiments, minimizing the time of data collection between the settings required for the algorithm. Since the results were satisfactory, the work corroborated the benefits of virtualization in the execution of this type of models with a large amount of calculations and a graphical component.
This paper proposes a fuzzy logic model about coffee borer propagation behavior at a colombian context, increasing the information about coffee borer propagation, beyond its growth (topic that has considerable information), looking forward to generate an impact taking into account the existing harm level generated by this plague on this important national product.
In this paper, we propose an ANFIS model based that verify their effectiveness compared to real data to determine the coffee berry borer proliferation risk in a colombian context, verifying its effectiveness compared to real data collected by the ICA in a study of the Integrated Management of Broca. As an initial step, four input variables are determined and a fuzzy model (made in another job), including factors were used: temperature, altitude, age and quality of the crop harvesting. A set of 144 rules and 20 times of training, the results were satisfactory and close to the actual infestation level, fulfilling the target of generating information regarding this problem by using engineering techniques.