A note on approximate Karush-Kuhn-Tucker conditions in locally Lipschitz multiobjective optimization

Nguyen Van Tuyen,Jen-Chih Yao,Ching-Feng Wen

DOI: https://doi.org/10.1007/s11590-018-1261-y

2018-04-13

Abstract:In the recent paper of Giorgi, Jiménez and Novo (J Optim Theory Appl 171:70--89, 2016), the authors introduced the so-called approximate Karush-Kuhn-Tucker (AKKT) condition for smooth multiobjective optimization problems and obtained some AKKT-type necessary optimality conditions and sufficient optimality conditions for weak efficient solutions of such a problem. In this note, we extend these optimality conditions to locally Lipschitz multiobjective optimization problems using Mordukhovich subdifferentials. Furthermore, we prove that, under some suitable additional conditions, an AKKT condition is also a KKT one.

Optimization and Control

What problem does this paper attempt to address?

### Problems Addressed by the Paper

This paper primarily explores how to obtain optimality conditions under approximate Karush-Kuhn-Tucker (AKKT) conditions in local Lipschitz multi-objective optimization problems. Specifically, the authors extend the AKKT conditions proposed by Giorgi, Jiménez, and Novo in 2016 for smooth multi-objective optimization problems to local Lipschitz multi-objective optimization problems, using Mordukhovich subdifferentials to achieve this goal.

### Main Contributions

1. **Extension of AKKT Conditions**: The authors extend the AKKT conditions from smooth multi-objective optimization problems to local Lipschitz multi-objective optimization problems, using Mordukhovich subdifferentials to handle non-smooth functions.
2. **Necessary and Sufficient Conditions**: It is proven that under certain suitable additional conditions, the AKKT conditions are also KKT conditions.
3. **Algorithm Design**: These results are very useful for designing algorithms to find approximately efficient solutions for non-smooth multi-objective optimization problems.

### Background and Motivation

- **KKT Conditions**: The KKT conditions are among the most important results in optimization theory, but they do not necessarily hold at local minima without satisfying certain constraint qualifications.
- **AKKT Conditions**: Giorgi, Jiménez, and Novo introduced the AKKT conditions to handle smooth multi-objective optimization problems and proved that these conditions hold for local weakly efficient solutions without additional requirements.
- **Non-Smooth Optimization**: This paper focuses on how to obtain similar AKKT conditions in local Lipschitz multi-objective optimization problems, which is significant in practical applications since many real-world problems are non-smooth.

### Conclusion

By introducing Mordukhovich subdifferentials, the authors successfully extend the AKKT conditions to local Lipschitz multi-objective optimization problems and prove that under certain conditions, the AKKT conditions are equivalent to the KKT conditions. These results provide a theoretical foundation for designing algorithms to solve non-smooth multi-objective optimization problems.