Hybridization of the Flower Pollination Algorithm—A Case Study in the Problem of Generating Healthy Nutritional Meals for Older Adults

Abstract

This chapter investigates the hybridization of the state-of-the-art Flower Pollination Algorithm as a solution for improving its execution time and fitness value in the context of generating healthy nutritional meals for older adults. The proposed hybridization approach replaces the local and global pollination operations from the Flower Pollination Algorithm with Path Relinking-based strategies aiming to improve the quality of the current solution according to the global optimal solution or to the best neighbouring solution. We model the problem of generating healthy nutritional meals as an optimization problem which aims to find the optimal or near-optimal combination of food packages provided by different food providers for each of the meals of a day such that the nutritional, price, delivery time and food diversity constraints are met. To analyse the benefits of hybridization, we have comparatively evaluated the state-of-the-art Flower Pollination Algorithm, adapted to our problem of generating menu recommendations, versus the hybridized algorithm variant. Experiments have been performed in the context of a food ordering system experimental prototype using a large knowledge base of food packages developed in-house according to food recipes and standard nutritional information.