Adaptive therapy is an approach to controlling therapeutic resistance in cancer that was inspired by integrated pest management. The central idea is to preserve and use chemosensitive cells to out-compete chemoresistant cells. However, there are multiple variations of adaptive therapy protocols and each protocol has multiple parameters. Furthermore, there is a combinatorial explosion of possible protocols and parameters when we start to investigate how to combine multiple drugs in adaptive therapy. We have used an agent-based model, based on the HAL library, to investigate how to optimize adaptive therapy protocols with either cytotoxic or cytostatic drugs. We have also investigated how to best prevent multidrug resistance when we combine two drugs in adaptive therapy protocols. We find that there is often a Goldilocks level of drug dosing that works best. If we use too much drug, the tumor quickly evolves resistance. However, if we use too little drug, we cannot even control the sensitive cells. In general, it is clear we should be using the minimum effective dose, which is standard in pest management. This even works for metronomic therapies that do not respond to tumor burden changes. Adaptive therapies work best when we change dosing as soon as we can detect a change in tumor burden, raise or lower the dose by at least 50% when we do change dose, and pause dosing as soon as the tumor shrinks appreciably. We also find that when utilizing more than one drug, we can prevent multidrug resistant clones from expanding if we only apply one drug at a time. We have tested a number of these protocols in a mouse model of hormone refractory ER+ breast cancer.