Recommender System Metaheuristic for Optimizing Decision-Making Computation

We implement a novel recommender system (RS) metaheuristic framework within nonlinear NP-hard decision-making problem, for reducing the solution search space before high-burden computational steps are performed. Our RS-based supports consideration of comprehensive evaluation criteria, including estimations potential set’s optimality, diversity, and feedback/preference end-user, while also being fully compatible with additional established RS metrics. Compared to prior Operations Research metaheuristics, our allows (1) achieving near-optimal scores through deep learning training, (2) fast parameter inference during instantiation trials, (3) ability reuse this trained module traditional ranking final options end-user. When implementing an experimental high-dimensionality simulation environment, we see average 91.7% reduction in computation time against baseline approach, 9.1% theoretical optimal scores. A simplified technique was developed more realistic environment dealing multidomain command control scenarios, where significant 87.5% is achieved compared maintaining 9.5%