Recommending Healthy Personalized Daily Menus—A Cuckoo Search-Based Hyper-Heuristic Approach

This paper presents a food marketplace-based system, which enables food providers to publish their food menus, and clients to order daily food menus personalized according to their profile. The proposed system integrates a Cuckoo Search based hyper-heuristic, which is a high-level method that selects and combines low-level heuristics in order to identify a sequence of low-level heuristics which lead to a menu for 1 day which best satisfies the profile of a client. In our approach, a daily food menu is composed of three main meals and two snacks and is generated by combining food menus for breakfast, lunch, dinner and snacks that are provided by various catering companies. As low-level heuristics, we considered random mutation (i.e. single-point/multiple-point mutation), random crossover (i.e. single-point/ multiple-point crossover) and memory-based mutation and crossover heuristics. We have evaluated the Cuckoo Search based hyper-heuristic on different client profiles and on 2600 menus.