| Resumo: | This dissertation aims the design and implementation of algorithms for solving an optimization problem of container capacity versus collection frequency, taking advantage of the information about the waste volume in containers obtained through sensors. Initially, state of the art will be made relative to intelligent cities through the use of sensors and the areas where internet of things is best applied. Waste collection optimization problems studied in the literature will be reviewed. Subsequently, through data provided by sensors that measure the volume occupied in each container, a statistical analysis is done on the same in order to understand which critical containers (whose capacity becomes null with some frequency) and which secondary containers (whose frequency of collection could be reduced). Next, algorithms are created to solve the problem of container capacity versus collection frequency, new models of frequency capacity are generated and results are present. Finally, the models generated are validated through prediction models and conclusions are drawn about the performance of each of the presented algorithms.
|
|---|