THE EFFECT OF DISCOUNTS ON THE COMPUTATION OF OPTIMAL MEDIA PLANS.

Abstract : Mathematical programming and heuristic programming have been frequently proposed to compute optimal media plans. While early efforts utilized a linear programming model, it is generally acknowledged that the response or effectiveness function to be maximized is non-linear. It has previously been shown that if the response surface is not well-behaved (i.e., not concave), it is possible for a local optimum found by mathematical programming or a heuristic approach not to be the global optimum. The paper shows that even if the response surface is concave, the existence of media discounts in the budget constraint is sufficient to cause these approaches to locate suboptimal solutions. Thus, additional heuristics are needed which are able to search for schedules that are not 'near' the current proposed schedule. In the final section, it is suggested that the media planning problem might be treated from a different viewpoint by eliminating the budget constraint altogether and working from basic economic principles. (Author)