Recommender systems in multi-stakeholder environments often need to optimize multiple objectives at the same time to meet the needs of sellers and consumers. Implementing the recommendations in these settings depends on properly combining the objectives to address the expectations of each participant, often with a defined and limited skin function. Actually, choosing these instruments becomes the next problem. Recent work has developed algorithms that adapt weights based on application-specific requirements by using RL to train the model. While this solves automatic weight synchronization, such methods do not work with standard weight synchronization. They don't allow human intervention too often dictated by business needs. To close this gap, we propose a multi-objective multi-objective recommendation framework that works with a small number of objectives. It also enables business decision makers to easily implement efficiency by assigning different values to multiple factors. We demonstrate the effectiveness and efficiency of our framework through improvements in online business metrics.
