ICACT20220022 Slide.26
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
Thank you for your attention
|
|
ICACT20220022 Slide.25
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
Reduce the cost for establishing the evolution algorithm. Based on cloud environment, computing ability can be increased. The proposed cloud based evolution algorithm can count the optimal solution for users via distributed system and parallel processing architecture.
|
|
ICACT20220022 Slide.24
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
After about 800 generations, the proposed evolution can find the optimal solution corresponding to the 1. multiple sets related to the best relationship of the disassembly tasks and 2. the fewest workstations for the disassembly tasks. which shown as follows.
|
|
ICACT20220022 Slide.23
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
One optimal total operation time could be corresponding to various operation sequences.
Based on the proposed algorithm, the various operation sequences with optimal operation time can be found.
The time cost for optimal solution searching is less than 10 seconds.
|
|
ICACT20220022 Slide.22
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
One optimal total operation time could be corresponding to various operation sequences.
Based on the proposed algorithm, the various operation sequences with optimal operation time can be found.
The time cost for optimal solution searching is less than 10 seconds.
|
|
ICACT20220022 Slide.21
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
Scenario 1: All the possible solutions corresponding to disassembly line tree could be found.
|
|
ICACT20220022 Slide.20
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
Cloud Computing: To enhance the computing performance, huge computing resource is needed.
The cloud environment based on the virtualization technologies is developed in this research.
The HBase, database on cloud environment, is proposed.
With the noSQL type database, each gene included in the chromosome can be recorded and changed dynamically.
Based on HBase on cloud, various length of different chromosomes could be individually recorded.
|
|
ICACT20220022 Slide.19
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
Evaluation and Fitness: Considering the disassembly line network tree structure , the relationship between different nodes or tasks can be defined.
Two relationship subjects are defined as invalid relationships:
No relationship: a task included in the chromosome is not the forward task and the following task of anther task, and also not the side branch task of any other tasks included in the same chromosome
Sequential relationship: According to the disassembly line network tree structure, the sequence of the tasks included in the same chromosome is not reversed.
|
|
ICACT20220022 Slide.18
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
Mutation: Since there are different methods for the crossover of chromosomes, suitable mutation methods are needed.
Uniform mutation: Inside one chromosome, one of the three attributes, Task, Operator, and Workstation, is randomly selected.
According to the on demand given probability value, one of two uniform mutation phases would be triggered: Exchanging Phase, and Replacement Phase.
The One-point mutation: Similar to the One-point crossover, a cut point will be randomly chosen.
According to the on demand given probability value, one of two One-point mutation phases would be triggered: Sequence Mutation and Variable Length Mutation.
|
|
ICACT20220022 Slide.17
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
In other words, at least one crossover method would be executed.
The task number, operator number, workstation number, and the total amount of the substrings or the length of the chromosome could be exchanged between parent chromosomes.
|
|
ICACT20220022 Slide.16
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
The flowchart of the hybrid crossover methods with the sequence
|
|
ICACT20220022 Slide.15
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
In this research, since there are three attributes such as Task, Operator, and Workstation, included in the chromosome.
To suitable crossover these three attributes and the various length based on the crossover rate is needed.
|
|
ICACT20220022 Slide.14
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
The operations of the proposed hybrid crossover method can be given as follows：
Uniform Crossover for genes corresponding to various attributes:
Based on the uniform crossover method, only the limited scope to the specific gene rather than multiple genes or whole chromosome.
In this research, the uniform crossover changes the number of task, or the operator number, or the workstation number individually.
One-point Crossover for substring exchanging:
The point which is located between substrings included in the same chromosome is randomly selected.
Since the point is randomly chosen, the total amount of exchanged substrings from different chromosomes could be also different.
|
|
ICACT20220022 Slide.13
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
Selection
Two chromosomes included in the current population of the current generation will be selected.
Considering the risk of local optimal solution searching and the performance of convergence, the differences between two selected chromosomes should be an important issue.
In this research, two available selection decisions could be made before the Roulette Wheel Selection method.
Crossover
The parallel distributed multiple phases crossover system including multiple phases crossover consideration and parallel distributed multiple crossover methods is proposed in this research.
|
|
ICACT20220022 Slide.12
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
The genetic algorithm cab be presented as follows：
Representation and Initialization
Each task of the binding node represented as the gene of the chromosome should be assigned an operator.
An operator can deal with single or multiple tasks at different workstations.
Each task, individual workstation, and the corresponding assigned operator can be combined as the working information set called substring, includes three genes: task number, assigned operator number, and the workstation number, included in the chromosome.
|
|
ICACT20220022 Slide.11
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
Based on the combined working set, the disassembly operation time of multiple working sets at workstation should subject to the Workstation Time Limitation.
The relationship between different working set should also subject to the Sequential Limitation.
To adopt the adaptive crossover method for enhancing the performance of optimal solution searching should be considered
|
|
ICACT20220022 Slide.10
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
To represent the possible solution as the chromosome of genetic algorithm, the variable length chromosome is used.
According to the total amount of the tasks corresponding to binding nodes, the same amount of operators will be assigned.
Each task of the individual binding node with an assigned operator is combined as a working set.
|
|
ICACT20220022 Slide.09
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
To solve the disassembly line balancing problem, Transformed AND/OR Graph (TAOG) was proposed to present the relationship or connection between different binding nodes.
|
|
ICACT20220022 Slide.08
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
There would be three possible disassembled partitions：
Remanufacturing for re-constructing new product
Recycling for re-using the product
Discarded
Three important factors are considered：
Workstations
Disassembly cell (binding node)
Disassembly line
|
|
ICACT20220022 Slide.07
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
The procedures of the genetic algorithm are introduced as follows:
Survival：Similar to the “survival of the fittest” rule from Darwin, the chromosome with better fitness value and within the population size can survive.
Checking：If the conditions for stopping the evolution algorithm are satisfied, the procedure for searching stops. Otherwise, the algorithm will start new generation from selection for new offspring chromosomes creation.
|
|
ICACT20220022 Slide.06
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
The main contributions of this research are:
Even there exist different sets or routes of the disassembly line network, the proposed algorithm can still find the optimal solution with all routes considered.
The optimal solution of the disassembly line network can be found by reviewing all available and found routes.
The convergence performance of the proposed genetic algorithm can be enhanced.
The optimal solution corresponding to the disassembly sequence of the disassembly line network can be found.
|
|
ICACT20220022 Slide.05
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
The main contributions of this research are:
Even there exist different sets or routes of the disassembly line network, the proposed algorithm can still find the optimal solution with all routes considered.
The optimal solution of the disassembly line network can be found by reviewing all available and found routes.
The convergence performance of the proposed genetic algorithm can be enhanced.
The optimal solution corresponding to the disassembly sequence of the disassembly line network can be found.
|
|
ICACT20220022 Slide.04
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
The main contributions of this research are:
Even there exist different sets or routes of the disassembly line network, the proposed algorithm can still find the optimal solution with all routes considered.
The optimal solution of the disassembly line network can be found by reviewing all available and found routes.
The convergence performance of the proposed genetic algorithm can be enhanced.
The optimal solution corresponding to the disassembly sequence of the disassembly line network can be found.
|
|
ICACT20220022 Slide.03
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
To find the optimal solution, evolution algorithm or genetic algorithms which based on the concept of “survival of the fittest” were proposed.
This research propose the cloud based modular genetic algorithm with variable length multiple crossover method for the optimal solution of the disassembly line network.
|
|
ICACT20220022 Slide.02
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
Recently, the environment protection is the popular issue in trade item production.
To minimize the material waste when recycling, to disassemble the objects with the on demand defined processes is the important issue.
How to find the suitable even optimal disassembling process with minimized time spending, and total numbers of disassembling machines and workers are important for various objects.
|
|
ICACT20220022 Slide.01
[Big slide for presentation]
[YouTube] |
Chrome Text-to-Speach Click!! |
 |
Today, the research Multi-Objective Optimization Based on Adaptive Evolution Algorithm for Disassembly Line Network Problems is presented.
I am Peng-Wen Wang, from National Formosa University, Taiwan
|
