Interactive selection of time-tables generated using evolutionary multi-objective optimization

Automatic generation of a time table with multi-level constraints is a challenging exercise. Typically, a timetable problem has many possible solutions in the initial search space, each with a distinct fitness level. In this paper we propose a multi-stage hybrid solution based on an evolutionary algorithm where the initial population is first generated that satisfies all the hard constraints. A user configurable fitness function is used to test the soft constraints. Solutions with fitness value above a certain threshold are then mutated to improve the fitness level of the next population and an acceptable optimal solution can then be selected interactively by the user. This unique interactive approach with configurable fitness functions allows more control to the user over the selection of an optimal solution and helps in multi-objective decision making.