Optimal Advertising Campaign Generation For Multiple Brands Using Multi-Objective Genetic Algorithm

The paper proposes a new modified multi-objective genetic algorithm for the problem of optimal TV advertising campaign generation for multiple products. This NP-hard combinatorial optimization problem with numerous constraints is one of the key issues for an advertising agency when producing the optimal television mediaplan. The classical approach to the solution of this problem is the greedy heuristic, which relies upon the strength of the preceding commercial breaks when selecting the next break to add to the campaign. While the greedy heuristic is capable of generating only \ group of solutions that are closely related in the objective space, the proposed modified multi-objective genetic algorithm produces a Pareto-optimal set of chromosomes that (i) outperform the greedy heuristic; and (ii) let the mediaplanner choose from a variety of uniformly distributed trade-off solutions. To achieve these results, the special problem-specific solution encoding, genetic operators and original local optimization routine were developed for the algorithm. These techniques allow the algorithm manipulating with only feasible individuals, thus significantly improving its performance that is complicated by the problem constraints. The efficiency of the developed optimization method is verified using the real data sets from the Canadian advertising industry.