Multi-objective Optimization: A Case Study

Nazmul Hasan

DOI: https://doi.org/10.48550/arXiv.2301.09166

2023-01-23

Abstract:The aim of this literature is to illustrate the application of multi-objective optimization routines through a case study of face milling operation. For this purpose, the face milling operation is designed as a multi-objective optimization problem and then solved to obtain optimum values for the machining parameters - cutting speed (Vc), feed rate (fz) and depth of cut (t) using the optimization routines. The formulated problem of face milling operation includes two conflicting objectives - to maximize Material Removal Rate (MRR) and to minimize surface roughness (Ra). Among various multi-objective optimization routines, five of them namely Global Criterion Method, Lexicographic Method, Weighted Sum Method, Epsilon Constraint Method and Genetic Algorithm are used in this literature. The outcomes of these multi-objective optimization routines are then compared to reflect their relative attractiveness.

Optimization and Control

What problem does this paper attempt to address?

The paper aims to address multi-objective optimization problems and demonstrates the application of multi-objective optimization methods through a specific case study—face milling operations. Specifically, the goal of the paper is to maximize the Material Removal Rate (MRR) while minimizing the Surface Roughness (Ra), and to select the optimal cutting parameters, namely cutting speed (Vc), feed rate (fz), and depth of cut (t), through an appropriate trade-off scheme. To achieve this goal, the paper compares five different multi-objective optimization methods:

1. **Global Criterion Method**: Finds the optimal solution by minimizing the differences between each objective function.
2. **Lexicographic Method**: Sorts the objective functions according to the decision maker's preference order and optimizes them sequentially.
3. **Weighted Sum Method**: Generates different Pareto optimal solutions by adjusting the weight factors.
4. **ε – Constraint Method**: Treats one objective function as the primary optimization target while considering other objectives as constraints.
5. **Genetic Algorithm**: Uses evolutionary algorithms to find the Pareto optimal set.

By comparing these five methods, the paper aims to identify the most applicable method under different circumstances, enabling decision-makers to choose the appropriate method based on actual conditions to solve the problem. Additionally, the paper compares the computational efficiency of different methods to evaluate their performance in practical applications.