Negotiation is the art of resolving conflicts of interests by finding outcomes that all parties agree with. Humans negotiate often, as there are many conflicts of interest in everyday life when working with other humans. Computer scientists have build computer agents that are capable of negotiating with each other in an attempt to outperform or replace humans in negotiation processes. These software-based agents are based on a negotiation strategy that is designed by a researcher. However, as for humans, creating an optimal negotiation strategy is difficult and sticking to the same strategy regardless of the negotiation setting is a bad idea (e.g. cultural differences). In this thesis, we automate the process of strategy configuration for automated negotiation agents to improve negotiation strategies with less human effort. In contrary to past attempts on this topic, the scale of the problem size and the applicability are vastly increased. We also automated the creation of multiple complementary strategies to exploit differences between negotiation settings and pick successful counter strategies. We show that we are capable of finding better strategies through automation and that switching between strategies based on the setting improves negotiation pay-off.