Modern manufacturing and logistics systems are supported by increasingly ubiquitous and powerful computing networks. Within these networks, oceans of data are continuously being generated by sensor...

Semiconductor manufacturing is a complicated flexible job-shop scheduling problem (FJSP) of combinatorial complexity. Because of the adoption of advanced process control and advanced equipment control, the processing time in advanced wafer fabs become uncertain. Existing approaches considering constant processing time may not be appropriate to address the present problem in a real setting. In practice, processing times can be represented as intervals with the most probable completion time somewhere near the middle of the interval. A fuzzy number that is a generalized interval can represent this processing time interval exactly and naturally. This paper developed a hybrid approach integrating a particle swarm optimization algorithm with a Cauchy distribution and genetic operators (HPSO+GA) for solving an FJSP by finding a job sequence that minimizes the makespan with uncertain processing time. In particular, the proposed hybridized HPSO+GA approach employs PSO for creating operation sequences and assigning the time and resources for each operation, and then uses genetic operators to update the particles for improving the solution. To estimate the validity of the proposed approaches, experiments were conducted to compare the proposed approach with conventional approaches. The results show the practical viability of this approach. This paper concludes with discussions of contributions and recommends directions for future research.

A semiconductor distributor that plays a third-party role in the supply chain will buy diverse components from different suppliers, warehouse and resell them to a number of electronics manufacturers with vendor-managed inventories, while suffering both risks of oversupply and shortage due to demand uncertainty. However, demand fluctuation and supply chain complexity are increasing due to shortening product life cycle in the consumer electronics era and long lead time for capacity expansion for high-tech manufacturing. Focusing realistic needs of a leading distributor for semiconductor components and modules, this study aims to construct a UNISON framework based on deep reinforcement learning (RL) for dynamically selecting the optimal demand forecast model for each of the products with the corresponding demand patterns to empower smart production for Industry 3.5. Deep RL that integrates deep learning architecture and RL algorithm can learn successful policies from the dynamic and complex real world. The reward function mechanism of deep RL can reduce negative impact of demand uncertainty. An empirical study was conducted for validation showing practical viability of the proposed approach. Indeed, the developed solution has been in real settings.

Wafer bin maps (WBM) in circuit probe (CP) tests that present specific defect patterns provide crucial information to identifying assignable causes in the semiconductor manufacturing process. However, most semiconductor companies rely on engineers using eyeball analysis to judge defect patterns, which is time-consuming and not reliable. Furthermore, the conventional statistical process control used in CP tests only monitors the mean or standard deviation of yield rates and failure percentages without detecting defect patterns. To fill the gap, this study aims to develop a manufacturing intelligence solution that integrates spatial statistics and neural networks for the detection and classification of WBM patterns to construct a system for online monitoring and visualisation of WBM failure percentages and corresponding spatial patterns with an extended statistical process control chart. An empirical study was conducted in a leading semiconductor company in Taiwan to validate the effectiveness of the proposed system. The results show its practical viability and thus the proposed solution has been implemented in this company.