Smart bins, equipped with sensors and IoT technologies, play a crucial role in optimizing waste collection by providing real-time data on bin fill levels. This paper introduces a Markovian Agent Model to simulate and evaluate different garbage collection strategies in a smart bin system. By analyzing various alarm thresholds and routing policies, the study identifies optimal approaches for minimizing overflows and enhancing collection efficiency. The results demonstrate that a strategy combining responsive alarm handling with route resumption (Resume policy) and a higher alarm threshold improves system stability and operational effectiveness.