A fuzzy evolving DL-XAI approach for explainable and trustworthy production planning in ubiquitous manufacturing systems

The centralized production planning mechanism of a ubiquitous manufacturing (UM) system is often deep and not transparent enough for participating manufacturers, thereby damaging their trust in the UM system. To deal with this issue, this study proposes a fuzzy evolving deep-learning (DL)-explainable artificial intelligence (XAI) approach. In the proposed fuzzy evolving DL-XAI approach, the selection of suitable manufacturers is treated as a fuzzy analytic hierarchy process (FAHP) problem, and its consistency is enhanced by optimizing a fuzzy mixed integer-nonlinear programming (FMINLP) model with the aid of a genetic algorithm (GA). The system administrator then distributes the quantity required for an order among multiple manufacturers for simultaneous production using the fuzzy technique for order preference by similarity to the ideal solution (FTOPSIS). A random forest (RF) is then built to explain the FAHP-FMINLP-GA-FTOPSIS mechanism and results. In comparison, most past XAI research has constructed linear regression (LR) equations to approximate deep neural networks (DNNs). The motivation for using a RF approximator for LIME is also explained in this study. Based on the RF, a local explanation can be generated for a specific manufacturer. A systematic procedure is also established to generate a contrastive explanation to provide a feasible improvement target that the manufacturer can pursue. Based on these explanations, the most important criteria for the manufacturer can be found. The proposed fuzzy evolving DL-XAI approach has been applied to the production planning of a UM system consisting of 3D printing facilities, in which the accuracy of fuzzy priority derivation was improved by 54%, ensuring the correctness of subsequent operations. In addition, the effectiveness of load balancing was also improved by 66%. Furthermore, the local and contrastive explanations generated from the RF were provided to manufacturers to help them understand and trust the operations of the UM system.

Authors

Institutions

Publication Details

Journal
Complex & Intelligent Systems
Published
2026-11-16
DOI
https://doi.org/10.1007/s40747-026-02577-w
Primary Topic
Explainable Artificial Intelligence (XAI)
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

A fuzzy evolving DL-XAI approach for explainable and trustworthy production planning in ubiquitous manufacturing systems

Toly Chen, Chin-Hau Huang, Yu-Cheng Wang, Yu-Cheng Lin
2 links
Complex & Intelligent Systems
Explainable Artificial Intelligence (XAI)
article

A fuzzy evolving DL-XAI approach for explainable and trustworthy production planning in ubiquitous manufacturing systems

Toly Chen, Chin-Hau Huang, Yu-Cheng Wang, Yu-Cheng Lin
article en

Abstract

The centralized production planning mechanism of a ubiquitous manufacturing (UM) system is often deep and not transparent enough for participating manufacturers, thereby damaging their trust in the UM system. To deal with this issue, this study proposes a fuzzy evolving deep-learning (DL)-explainable artificial intelligence (XAI) approach. In the proposed fuzzy evolving DL-XAI approach, the selection of suitable manufacturers is treated as a fuzzy analytic hierarchy process (FAHP) problem, and its consistency is enhanced by optimizing a fuzzy mixed integer-nonlinear programming (FMINLP) model with the aid of a genetic algorithm (GA). The system administrator then distributes the quantity required for an order among multiple manufacturers for simultaneous production using the fuzzy technique for order preference by similarity to the ideal solution (FTOPSIS). A random forest (RF) is then built to explain the FAHP-FMINLP-GA-FTOPSIS mechanism and results. In comparison, most past XAI research has constructed linear regression (LR) equations to approximate deep neural networks (DNNs). The motivation for using a RF approximator for LIME is also explained in this study. Based on the RF, a local explanation can be generated for a specific manufacturer. A systematic procedure is also established to generate a contrastive explanation to provide a feasible improvement target that the manufacturer can pursue. Based on these explanations, the most important criteria for the manufacturer can be found. The proposed fuzzy evolving DL-XAI approach has been applied to the production planning of a UM system consisting of 3D printing facilities, in which the accuracy of fuzzy priority derivation was improved by 54%, ensuring the correctness of subsequent operations. In addition, the effectiveness of load balancing was also improved by 66%. Furthermore, the local and contrastive explanations generated from the RF were provided to manufacturers to help them understand and trust the operations of the UM system.

Complex & Intelligent Systems
Chaoyang University of Technology (TW), National Yang Ming Chiao Tung University (TW), National Chin-Yi University of Technology (TW)
Industry, innovation and infrastructure
Explainable Artificial Intelligence (XAI)
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon