A Multiobjective Approach for E-Commerce Website Structure Optimization

Abstract

Complex websites comprise a variety of diverse web entities, which require constant restructuring resonating with the latest trends, shifting consumer expectations and market driven changes. Therefore, designing suitable models to optimally restructure such websites is of paramount importance and must take into consideration several factors about the web entities such as display size, download time, type, location in the page, sales likelihood, discounts, and the ongoing trend. A recent study has taken all these attributes into consideration and designed a model based on the Access Score, Interface Score, and Purchase Score. However, this model suffers from certain drawbacks such as it did not address the underlying cohesiveness between these attributes. Further, it provided a single optimal solution to the adaptive website structure optimization (AWSO) problem and relied on the a priori knowledge of weights. The basis of the new proposed model is that there can be more than one optimal solution to the AWSO problem in the real world. The novel tri-objective optimization model uses NSGA-II algorithm to simultaneously optimize the attributes and finds advantageous trade-off solutions without requiring a priori knowledge of weights. The proposed MO-AWSO_NSGA-II model is shown to outperform the existing model proving it better suited for the AWSO problem.