From a medical point of view the problem in drug discovery is to find a compound that causes a desired outcome, not one that interacts with a specific target. For instance, if you want to reduce mortality due to infection with Covid, you have many strategies that may be pursued at the molecular level, from vaccination to interference with viral propagation to reduction of symptoms or boosting of the immune response.
The difference between a combinatorial strategy and a rational strategy is to a first approximation simple: in a combinatorial strategy compounds are tested for activity without immediate concern for how they bring about an outcome. There is no "a priori" information to guide compound selection other than the common properties of known drugs (as encoded eg by Lipinski's rules) and the ability to procure or synthesize compounds. Therefore the "rational" approach in this case is to create as large a library and to screen as many compounds as is reasonable given cost and time.
Rational drug design takes into consideration a molecular target (such as a protein) and perhaps the mechanism of action of a known lead compound and designs future potential drug compounds around that target and/or lead. That narrows the set of compounds under consideration reducing time and cost.
In general, the strategies of combinatorial and rational design are not mutually exclusive. It is typical to start a rational strategy after early leads have been identified some other way, say through combinatorial chemistry. It is also common to design a set of compounds during a stage in rational drug design using a combinatorial strategy applied to develop next generation leads from a starting compound.