Objective Optimization
Location
Ada, Ohio
Start Date
3-12-2024 12:00 AM
End Date
3-12-2024 12:00 AM
Description
The research I conducted over this semester was on the investigation of single-objective evolutionary optimization of the maximum clique problem. The maximum clique problem in question, is a problem that attempts to find the largest number edges within a subset of vertices that are all adjacent to one another in a graph. This problem being NP-complete requires exponential time to complete when more vertices are added to the graph. To attain this goal of optimizing the problem the study used the pymoo library and its documentation to amend sample problems into that of the maximum clique problem. To accomplish this, a definition of the problem and constraints were written in order to realize results of the simulated product. This problem definition and constraints were fed through a genetic algorithm where it takes the best one hundred results out of the two hundred generated solutions and uses them to breed the next generation. This process is then repeated until as many times as specified by the user. Through this a pattern of the algorithm progressively getting closer to the correct answer is observed. Although my preliminary results suggest that the algorithm is not finely tuned to the degree that would be appropriate for conclusivity it demonstrates convergence towards the optimal solution. To accomplish better results would mean to fine tune the constraints and the goals the algorithm is trying to achieve. Through this research it provides a framework towards further evolutionary optimization that could be done to other problems such as those that are multi-objective.
Recommended Citation
Kacir, Alex, "Objective Optimization" (2024). College of Engineering Student Research Colloquium. 10.
https://digitalcommons.onu.edu/eng_student_research_colloquium/2024/Presentations/10
Objective Optimization
Ada, Ohio
The research I conducted over this semester was on the investigation of single-objective evolutionary optimization of the maximum clique problem. The maximum clique problem in question, is a problem that attempts to find the largest number edges within a subset of vertices that are all adjacent to one another in a graph. This problem being NP-complete requires exponential time to complete when more vertices are added to the graph. To attain this goal of optimizing the problem the study used the pymoo library and its documentation to amend sample problems into that of the maximum clique problem. To accomplish this, a definition of the problem and constraints were written in order to realize results of the simulated product. This problem definition and constraints were fed through a genetic algorithm where it takes the best one hundred results out of the two hundred generated solutions and uses them to breed the next generation. This process is then repeated until as many times as specified by the user. Through this a pattern of the algorithm progressively getting closer to the correct answer is observed. Although my preliminary results suggest that the algorithm is not finely tuned to the degree that would be appropriate for conclusivity it demonstrates convergence towards the optimal solution. To accomplish better results would mean to fine tune the constraints and the goals the algorithm is trying to achieve. Through this research it provides a framework towards further evolutionary optimization that could be done to other problems such as those that are multi-objective.