Design of efficient algorithm for waste management /
Maintaining current municipal solid waste management (MSWM) for the next ten years would not be efficient anymore due to waste overflow in relation to the way of handling waste collection. This practice has brought too many environmental and health issues such as air pollution and typhoid fever resp...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
Kuala Lumpur :
Kulliyyah of Engineering, International Islamic University Malaysia,
2018
|
| Subjects: | |
| Online Access: | http://studentrepo.iium.edu.my/handle/123456789/4509 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Maintaining current municipal solid waste management (MSWM) for the next ten years would not be efficient anymore due to waste overflow in relation to the way of handling waste collection. This practice has brought too many environmental and health issues such as air pollution and typhoid fever respectively. Therefore, this project has proposed two Artificial Neural Network (ANN) based prediction algorithms that can forecast future Solid Waste Generation (SWG) based on two factors; population growth and household size, which can improve MSWM in Malaysia. For the population growth factor, the project can deduct that the SWG based on human population will always increase linearly with respect to time. However, the SWG based on household size would not increase linearly with time. Hence, two prediction algorithms are needed for the two factors. An online survey has been conducted to observe the human behaviour which motivates this research. Also, a smart waste bin has been developed that can measure the weight, detect the emptiness level of the waste bin, stores information and have direct communication between waste bin and collector crews. Collection of data for the prediction of SWG based on population growth factor uses the Malaysian population as sample size and the data is acquired via authorized Malaysia statistics' websites. Whilst, Kulliyyah of Engineering (KOE) in International Islamic University Malaysia (IIUM) has been chosen as the sample size for household size factor. All data will be normalized in the pre-processing stage before proceeding to the prediction using Visual Gene Developer. Statistical measure that was used in this project to evaluate the performance was the R2 value. For population growth factor, two hidden layers with ten nodes and five nodes in the first and second layers were used respectively. Whilst, for household size factor, two hidden layers were used and the number of nodes for the first and second layers were five and ten each. After the prediction is done, the result portrayed that there will be an increase of 29.03 percent of SWG in year 2031 compared to year 2012. Whilst, for household size factor, the average rate of increment is only 2.05 percent from week one until week twenty. The limitation to this study is that the data for population growth factor is not based on real time as it is restricted by the government. |
|---|---|
| Physical Description: | xiv, 105 leaves : colour illustrations ; 30cm. |
| Bibliography: | Includes bibliographical references (leaves 91-95). |