Production quantity estimation using an improved artificial neural network
By considering on the competitive market today, managing inventory becomes one factor that affected in improving business performance. This encouraged most industries to manage it efficiently by determining effective decision for inventory replenishment. For instance, mostly, industries decide next...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English |
| Published: |
2015
|
| Subjects: | |
| Online Access: | http://eprints.utem.edu.my/id/eprint/15868/1/Raden%20Nur%20Rachman%20Dzakiyullah.pdf http://eprints.utem.edu.my/id/eprint/15868/2/Production%20quantity%20estimation%20using%20an%20improved%20artificial%20neural%20network.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | By considering on the competitive market today, managing inventory becomes one factor that affected in improving business performance. This encouraged most industries to manage it efficiently by determining effective decision for inventory replenishment. For instance, mostly, industries decide next inventory replenishment by considering on their last historical production. However, this decision cannot be implemented on the next production due to uncertainty/fluctuated condition. Therefore, poor decision on producing product will influence the business’ costs. Hence, this research proposes model based on Neural Network Back Propagation (NNBP) to estimate production quantity. This model is designed based on input variables that affect the determination of production quantity which include demand, setup costs, production, material costs, holding costs, transportation costs. The performance of NNBP can be analyzed using Root Mean Square Error (RMSE) and Mean Absolute Error (MAE). In order to increase the performance of NNBP, optimization techniques such as Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) are being hybrid with the ANN model to become Hybrid Neural Network Genetic Algorithm (HNNGA) model and Hybrid Neural Network Particle Swarm Optimization (HNNPSO) model respectively. These techniques were used to optimize attribute weighting on NNBP model. The proposed models were examined using private dataset that collected from Iron Casting Manufacturing in Klaten, Indonesia. Moreover, validation is conducted for all proposed models through both Cross-Validation and statistical analysis. The cross-validation is common technique used to prevent over fitting problem by dividing the data into two categories namely data training and data test. Meanwhile, statistical analysis considers normality test on error estimation and the significant difference among the proposed models. Experimental result shows that HNNGA and HNNPSO provide smaller measurement error that concurrently improves the performance of NNBP model. In this work, the proposed model contributes not only to update the original instrument, but also applicable and beneficial for industry, particularly in deciding effective inventory replenishment decision on production quantity. |
|---|