Optimizing Online Promotion Planning: A Multi-Objective, Multi-Market, Multi-Period Approach

Abstract

This study focuses on retail promotion planning problem in a multi-objective, multimarket and multi-period framework and concomitantly addresses the following questions: What are the conditions for a retailer to offer an online promotion to all or a select set of the markets it is operating in? How should the retailer select the partial set of markets to offer a promotion? How should the selection of promotion period and markets be coordinated? Starting from a simple two market, single-period scenario, we gradually build our model to address the large scale promotion coordination problem via a multi-objective evolutionary algorithm with decomposition and pareto local search" (MOEA/D-PLS). The proposed algorithm allows retailers to coordinate online price promotions in multiple retail markets over multiple periods, minimizing demand leakage from offline to online channels and preventing demand drain and stock-outs. It decomposes the complex multi-objective optimization problem into a set of single-objective optimization problems, and uses an evolutionary algorithm to solve each problem simultaneously, improving the quality of solutions via a problem-specific Pareto local search method. This method allows large scale promotion planning problem, which is computationally non-trivial, to be solved optimally in an efficient manner. To demonstrate the proposed methodology, we discuss a numerical implementation of the algorithm using data from a large nation-wide pizza chain.

Keywords: Optimization models, mixed integer programming, omni-channel retailing, promotional planning