Journal
NEURAL COMPUTING & APPLICATIONS
Volume 33, Issue 8, Pages 3503-3519Publisher
SPRINGER LONDON LTD
DOI: 10.1007/s00521-020-05197-8
Keywords
Rock size distribution; Rock fragmentation; FFA-BGAM; Blasting; Optimization algorithm; Hybrid technique
Categories
Funding
- Center for Mining, Electro-Mechanical research of Hanoi University of Mining and Geology (HUMG), Hanoi, Vietnam
- National Natural Science Foundation Project of China [41807259]
- Innovation-Driven Project of Central South University [2020CX040]
- research team of Innovations for Sustainable and Responsible Mining (ISRM) of HUMG
Ask authors/readers for more resources
This paper introduces a new soft computing model FFA-BGAM for accurately modeling rock fragmentation using a firefly algorithm to optimize the BGAM model. Compared to other soft computing models, FFA-BGAM shows higher accuracy in predicting rock size distribution.
This paper proposes a new soft computing model (artificial intelligence model) for modeling rock fragmentation (i.e., the size distribution of rock (SDR)) with high accuracy, based on a boosted generalized additive model (BGAM) and a firefly algorithm (FFA), called FFA-BGAM. Accordingly, the FFA was used as a robust optimization algorithm/meta-heuristic algorithm to optimize the BGAM model. A split-desktop environment was used to analyze and calculate the size of rock from 136 images, which were captured from 136 blasts. To this end, blast designs were collected and extracted as the input parameters. Subsequently, the proposed FFA-BGAM model was evaluated and compared through previous well-developed soft computing models, such as FFA-ANN (artificial neural network), FFA-ANFIS (adaptive neuro-fuzzy inference system), support vector machine (SVM), Gaussian process regression (GPR), and k-nearest neighbors (KNN) based on three performance indicators (MAE, RMSE, andR(2)). The results indicated that the new intelligent technique (i.e., FFA-BGAM) provided the highest accuracy in predicting the SDR with an MAE of 0.920, RMSE of 1.213, andR(2)of 0.980. In contrast, the remaining models (i.e., FFA-ANN, FFA-ANFIS, SVM, GPR, and KNN) yielded lower accuracies in predicting the SDR, i.e., MAEs of 1.248, 1.661, 1.096, 1.573, 1.237; RMSEs of 1.598, 2.068, 1.402, 2.137, 1.717; andR(2)of 0.967, 0.968, 0.972, 0.940, 0.963, respectively.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available