A Decision Support Tool for Food Fraud Regulation: Hybrid Framework Combining Game Theory, DCA Optimization, and Machine Learning
Preprint
Abstract
This study presents a decision-support tool designed for the regulation of food fraud, based on on hybridizing of four approaches: i) an economic model inspired by the Chinese milk crisis, opposing a food firm and a public authority. The model integrates the main hypotheses of food fraud mechanisms, considering organoleptic standards, production and compliance costs, the firm's ability to substitute the required nutrient with a less expensive chemical substance, as well as the cost of acquiring that substance. It also considers the probability of a health crisis occurring, the financial penalty stipulated by regulations, the overall health risk, and the expected profits based on the firm's different strategic choices; ii) A sequential game where the authority first chooses the nutritional standard and the fine applicable in case of a health crisis to minimize health risk, and then the firm determines its optimal compliance level to maximize its expected profit. The game is solved by backward induction; iii) the DCA method to locally solve step 2 of the game, and iv) The ML model (CaatBoost) to estimate the firm's best response function for solving step 1 of the game. The experimental results of the tool demonstrate that the organoleptic standard primarily drives firm's compliance. Strict nutritional standards are effective to reduce health risk only when paired with high fines. Optimal regulation is adaptive: when production capacity, prices, or organoleptic standards are low, maximum standards and fines are required; otherwise,relaxing the standard best limits health risks.Keywords
All Related Versions
This publication has 19 references indexed in Scilit:
- Advanced detection tools in food fraud: A systematic review for holistic and rational detection method based on research and patentsFood Chemistry, 2024
- Adulteration involving starch and its evaluation methods2022
- Food Adulteration Detection using Artificial Intelligence: A Systematic ReviewArchives of Computational Methods in Engineering, 2021
- A game theoretic decision model for organic food supplier evaluation in the global supply chainsJournal of Cleaner Production, 2020
- Food fraud prevention shifts the food risk focus to vulnerabilityTrends in Food Science & Technology, 2017
- Food fraud: policy and food chainCurrent Opinion in Food Science, 2016
- Melamine in milk products in China: Examining the factors that led to deliberate use of the contaminantFood Policy, 2010
- Melamine-Contaminated Powdered Formula and Urolithiasis in Young ChildrenNew England Journal of Medicine, 2009
- The DC (Difference of Convex Functions) Programming and DCA Revisited with DC Models of Real World Nonconvex Optimization ProblemsAnnals of Operations Research, 2005
- Monopoly and product qualityJournal of Economic Theory, 1978