We will present contributions to two problems in CSPs: the first one is to understand better the computational hardness of problems, and the second one is to create tools to build, as desired, sets of problems with selected hardness and some predefined characteristics.
On the one hand, we will discuss several factors that contribute to the hardness in the typical case, and how hardness can be increased by modifying these factors. The problems analyzed are: Random Binary CSP, generalised Sudoku Problems and Edge Matching Puzzles. On the other hand, we will describe the mechanisms we have designed for building problem instances with adjustable hardness in a systematic and easy way.