Motivated by identifying equations to automate the discovery of conjectures about sharp bounds on combinatorial objects, we introduce a CP model to acquire Boolean-arithmetic equations (BAE) from a table providing sharp bounds for various combinations of parameters. Boolean-arithmetic expressions consist of simple arithmetic conditions (SAC) connected by a single commutative operator such as ‘∧’, ‘∨’, ‘⊕’ or ‘+’. Each SAC can use up to three variables, two coefficients, and an arithmetic function such as ‘+’, ‘−’, ‘×’, ‘floor’, ‘mod’ or ‘min’. We enhance our CP model in the following way to limit the search space: (i) We break the symmetries linked to multiple instances of similar SACs in the same expression. (ii) We prevent the creation of SAC that could be simplified away. We identify several use cases of our CP model for acquiring BAE and show its applicability for learning sharp bounds for eight types of combinatorial objects as digraphs, forests, and partitions.