DEVELOPMENT OF GENETIC ALGORITHM FOR ASSEMBLY LINE BALANCING WITH HUMAN-ROBOT COLLABORATION

Collaborative robot (cobot) is a new innovative robot designed to work side by side with humans and considered human safety aspects. The implementation of cobots in the industry is increasing in Indonesia for various production applications. Research on the Assembly Line Balancing Problem Human-Robo...

Full description

Saved in:
Bibliographic Details
Main Author: Novita Sitorus, Jessica
Format: Final Project
Language:Indonesia
Subjects:
Online Access:https://digilib.itb.ac.id/gdl/view/79330
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Institut Teknologi Bandung
Language: Indonesia
Description
Summary:Collaborative robot (cobot) is a new innovative robot designed to work side by side with humans and considered human safety aspects. The implementation of cobots in the industry is increasing in Indonesia for various production applications. Research on the Assembly Line Balancing Problem Human-Robot Collaboration (ALBP-HRC) has been conducted. The study proposed an analytical model. The weakness of the analytical model is the long computational time to find an optimal solution for data with a total of 35 and 45 tasks. Therefore, the analytical method is viable for small cases only. Thus, this research proposes a metaheuristic method to overcome previous research limitations. This research proposes the genetic algorithm as the metaheuristic method. In general, the algorithm contains two procedures: solution construction and solution improvement. Developments are in the area of generating feasible solutions, crossover, and mutation for ALBP-HRC cases. The proposed algorithm is translated to the Python programming language to execute the solutions. Based on the results, the proposed algorithm, in several cases, can match the optimal solution. Overall, the proposed algorithm achieved better solutions with an average solution gap of 13%. The computation time is faster, with an average time gap of 75%. The experimental designs also show that population and generation parameters significantly affect the performance of the genetic algorithm in finding solutions.