Modelling and Solving the Generalised Balanced Academic Curriculum Problem with Heterogeneous Classes

The Balanced Academic Curriculum Problem (BACP) consists in assigning courses to teaching periods satisfying prerequisites and balancing students’ load in terms of credits and number of courses. The BACP planning horizon is divided in academic years, and each academic year is divided into terms. Each term of a year is a teaching period in which courses can take place. The problem consists in finding an assignment of courses to periods that satisfies certain load limits and prerequisites. The first formulation of BACP has been proposed by Castro and Manzano [2], and it has been included in CSPLib [5, prob. 30] along with three benchmark instances. BACP, in the CSPLib formulation, has been studied by Hnich et al [6], Castro et al [1], Lambert et al [7], and Monette et al [8]. However, BACP is a very simplified model of the real problem that universities have to solve in practice. In fact, in its formulation it is implicitly assumed that a student takes all delivered courses without personal choices; whereas in practice a student can select among alternatives. To try to overcome such limitation and to deal with the real cases, the formulation has been extended by Di Gaspero and Schaerf [4]. The problem defined by Di Gaspero and Schaerf [4], called GBACP (G for Generalised), makes it possible to specify several curricula, with courses shared among them. A curriculum is a set of courses representing a possible complete selection of a student. In this new formulation, courses have to be balanced for each single curriculum. Moreover, GBACP includes professor’s preferences for teaching in specific terms, which are often taken into account in real situations.