A Multi-Objective Bayesian Approach with Dynamic Optimization (MOBADO). A Hybrid of Decision Theory and Machine Learning Applied to Customs Fraud Control in Spain

This paper studies the economically significant problem of optimization customs fraud control, which is a critical issue for many countries. The European Union (EU) alone handles 4693 tons goods every minute (2018 figures). Even though 70% are imported at zero tariff, EU raised EUR 25.4 billions in 2018, and customs-related income transferred by member states to accounts nearly 13% its overall budget. In this field, (a) conflicting objectives qualitative cannot be reduced common measure (security terrorism, health, drug market access taxes, etc.); (b) each submitted item has dozens characteristics; (c) there constraints; (d) risk analysis systems have make decisions real time. Although World Customs Organization promoted use artificial intelligence increase precision controls, very complex due data characteristics interpretability, requirement established officers. paper, we propose new Bayesian-based hybrid approach combining machine learning multi-objective linear programming (MOLP), called Bayesian with dynamic (MOBADO). We demonstrate that it possible more than double (with 237% increase) current inspection systems, freeing up almost 50% human resources, outperform past results respect above objectives. MOBADO an technique could combined any capable optimizing quality