Global Five-Continent Optimization of Personalized Nutrition and Exercise in Parkinson’s Disease Using Open-Access Data, Genetic Algorithms, and Fuzzy Logic

Abstract

Background: Parkinson’s disease (PD) is a heterogeneous neurodegenerative disorder in which nutritional status, physical activity, demographic characteristics, socioeconomic conditions, and genetic factors may contribute to differences in disease burden and intervention requirements. Conventional approaches generally provide standardized recommendations and may not adequately reflect geographical and population-level heterogeneity.

Objective: This study aimed to develop a computational framework for optimizing personalized nutrition and exercise profiles for Parkinson’s disease across five continents using open-access global datasets, fuzzy logic, and genetic algorithm optimization.

Methods: A retrospective computational framework covering the period from 1990 to 2026 was designed using freely accessible data on Parkinson’s disease burden, nutrition, physical activity, demographics, socioeconomic characteristics, and population genetics. Data were harmonized according to country, year, demographic structure, and measurement units. Nutritional and physical activity indicators were standardized and incorporated into a fuzzy inference system. Genetic information was used at the population level for genetic stratification rather than individual clinical prediction. A genetic algorithm was subsequently applied to optimize combinations of nutritional and exercise parameters using a multi-objective fitness function incorporating nutritional adequacy, exercise suitability, feasibility, safety, population compatibility, and risk reduction. Sensitivity analysis and repeated optimization runs were used to assess model stability.

Results: The proposed framework demonstrated substantial geographical heterogeneity in nutritional profiles, physical inactivity, Parkinson’s disease burden, and optimization scores. The illustrative optimization results showed overall scores of 0.781 for Africa, 0.816 for Asia, 0.854 for Europe, 0.839 for the Americas, and 0.872 for Oceania. The fuzzy inference system generated graded intervention priorities rather than binary classifications, while the genetic algorithm progressively improved candidate solutions across generations. Multi-objective optimization produced several Pareto-optimal nutrition–exercise profiles, indicating that a single universal intervention may not be appropriate for all populations.

Conclusion: The integration of open-access global data, fuzzy logic, and genetic algorithms provides a potentially reproducible computational framework for geographically adaptive nutrition and exercise optimization in Parkinson’s disease. The proposed approach may support hypothesis generation and decision-support research by integrating multiple sources of heterogeneity within a unified model. However, computationally optimized profiles should not be interpreted as clinical prescriptions or evidence of therapeutic efficacy. Prospective clinical validation using individual-level nutritional, functional, clinical, and genetic data is required before implementation in patient care.