A MODIFIED GENETIC ALGORITHM FOR MIXED MODEL ASSEMBLY LINE BALANCING PROBLEM EQUIPPED HUMAN-ROBOT COLLABORATION RESOURCES
The development of robot technology, namely collaborative robots (cobot), is able to assist human tasks in an assembly line. Previous research on the usage of cobot in assembly lines conducted using an analytical model has shown a weakness, which is the inability of the model to accommodate a relati...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/68849 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | The development of robot technology, namely collaborative robots (cobot), is able to assist human tasks in an assembly line. Previous research on the usage of cobot in assembly lines conducted using an analytical model has shown a weakness, which is the inability of the model to accommodate a relatively large number of assembly tasks in efficient computational time. In addition, the mixed-model assembly line with human-robot collaboration has the potential to be developed to adapt to product diversity and shorter product life.
One of the population-based metaheuristic methods, namely genetic algorithm, is developed to answer the shortcomings of the previous research and consists of several main procedures, which are initial population generation through a construction algorithm, tournament selection to select parent individuals from the population to be improved, crossover to form new and better individuals based on the parents’ characteristics, and mutation to help the process out of the local optimal. The objective function used in this research is to minimize the total assembly line cost. Python programming language is utilized to develop the algorithm.
Based on the analytical method, the genetic algorithm is able to produce optimal solutions for data of 11 to 19 tasks and feasible solutions for data of 25 to 43 tasks with a solution gap of 21,82% and a better computational time gap of 88,67%. Meanwhile, better solutions with a gap of 16,9% for big data of 61 and 75 tasks are found compared to the local optimal solutions from the analytical method. Based on the experiment, parameters that are significant in minimizing the objective function in efficient computation time are the number of generations (G) and the probability of mutation (????????). |
|---|