Using coevolution and substitution of the fittest for health and well-being recommender systems

TL;DR

This paper introduces substitution of the fittest (SF), a domain-independent technique for coevolutionary algorithms, which improves engagement and recommendation quality in health and well-being systems.

Contribution

The paper presents SF as a novel, calibration-free method that enhances engagement and solution quality in coevolutionary recommender systems, validated through both toy and real-world experiments.

Findings

SF maintains engagement better than existing techniques

Recommendations using SF are higher quality and more diverse

SF outperforms in both toy and real-world domains

Abstract

This research explores substitution of the fittest (SF), a technique designed to counteract the problem of disengagement in two-population competitive coevolutionary genetic algorithms. SF is domain-independent and requires no calibration. We first perform a controlled comparative evaluation of SF's ability to maintain engagement and discover optimal solutions in a minimal toy domain. Experimental results demonstrate that SF is able to maintain engagement better than other techniques in the literature. We then address the more complex real-world problem of evolving recommendations for health and well-being. We introduce a coevolutionary extension of EvoRecSys, a previously published evolutionary recommender system. We demonstrate that SF is able to maintain engagement better than other techniques in the literature, and the resultant recommendations using SF are higher quality and more diverse than those produced by EvoRecSys.