מאיר קלך

Many democratically elected representative governments around the world require the formation of a coalition to create a government that is expected to be durable and stable. Creating such a coalition is far from trivial. This research describes a computerized coalition formation decision support system that attempts to meet this challenging task and formally rank possible coalitions. We use theories from political science and utilize insights from artificial intelligence to build the tool. We demonstrate the tool’s utility and validity empirically by comparing historical coalitions to predictions made by the tool. Tests show that we are able to recommend quite accurately durable coalitions and assess which coalitions would most likely be unstable.