Waste collection is nowadays an increasingly important business. However, it is often an inefficient operation due to the high uncertainty associated with the real waste bins' fill-levels. To deal with such uncertainty the use of sensors to transmit real time information is seen as possible solution. But, in order to improve operations' efficiency, the sensors' usage must be combined with optimization procedures that inform on the optimal collection routes to operationalize, so as to guarantee a maximization of the waste collected while also minimizing transportation costs. The present work explores this challenge and studies three operational management approaches to define dynamic optimal routes, considering the access to real-time information on the bins' fill-levels. A real case study is solved and important results were found where significant profit improvements are observed when compared to the real operation. This shows the potential of the proposed approaches to build an expert system, which can support the operations manager's decisions.