An estimation of distribution algorithm with clustering for scenario-based robust financial optimization

Area of Science:

• Financial mathematics
• Optimization algorithms
• Risk management

Background:

• Financial optimization seeks robust investment plans balancing return and risk.
• Scenario-based uncertainty, where market conditions significantly impact investment performance, is under-explored in multi-objective optimization.
• Robust financial planning is crucial for maximizing returns while minimizing risk.

Purpose of the Study:

• To propose a novel algorithm for scenario-based robust financial optimization.
• To address the under-explored domain of scenario-based uncertainty in multi-objective optimization problems.
• To evaluate the algorithm's effectiveness on a real-world financial problem.

Main Methods:

• A nondominated sorting estimation of distribution algorithm with clustering (NSEDA-C) was developed.
• A simplified simulation method was used to measure investment return.
• An estimation model was devised to quantify investment risk.
• The NSEDA-C was applied to a robust group insurance portfolio problem.

Main Results:

• The proposed NSEDA-C algorithm effectively handles scenario-based uncertainty in financial optimization.
• The algorithm demonstrated its capability to balance investment returns and risks.
• Validation was achieved through application to a group insurance portfolio problem with real-world insurance products.

Conclusions:

• The NSEDA-C is an effective algorithm for solving scenario-based robust financial optimization problems.
• The study highlights the importance of considering scenario-based uncertainty in financial planning.
• The proposed method offers a valuable tool for developing robust investment strategies in insurance and finance.