A Decision Support System for self-promotion planning at a TV station

Abstract

This paper describes a Decision Support System (DSS) to plan and maintain the best assignment of the weekly self-promotion space for a TV station. The DSS allows the decision makers to generate, evaluate, and compare different solutions, as well as include additional knowledge by changing manually the solutions provided. The DSS comprises an input data base, a statistic module, a heuristic tool, and a graphical interface. The information used is real data provided by a Portuguese TV station and by a company that provides daily audience information based on the audience of previous days. The heuristic comprises two steps. Firstly, feasible solutions are provided by a greedy constructive heuristic, which are then improved by a genetic algorithm optimizer. Finally, the best solution found is presented as a grid of spots-breaks allocation, that is the spots to be broadcasted in each break for each day of the week. The interface can also be used to manually change the current solution, if desirable. The changed solution is then evaluated. The self-promotion plan comprises the allocation of the several self-promotion spots to the different time slots of the self-promotion breaks for the following seven days. However, the audience information is received daily, which permits to replan previous allocations for the next seven days in a receding horizon control strategy.