With the advent of Industry 4.0, Cyber Physical Systems (CPS) and Internet of Things (IoT) technology provide enormous opportunities and support for intelligent manufacturing. Material handling is vital in manufacturing systems to ensure that proper materials with the right quantity and quality can be delivered to each machine or workstation at the right time. AGV has been widely used in smart factories for material handling, and AGV scheduling plays a critical role in practical AGV application. However, the AGV scheduling problem becomes more and more complex with the development of intelligent manufacturing, making the design of effective and efficient scheduling algorithms complicated. In this paper, we analyzed the AGV dispatching tasks in the workshop and model the workshop as a node network, and apply an improved Multi-Agent Reinforcement Learning (MARL), that is, an improved QMIX model to solve the AGV scheduling problem. The experiment results show that the proposed approach outperforms the other commonly-used methods such as deep reinforcement learning (DQN) under different environment settings, in term of the maximum makespan of AGV.

Most enterprises concentrate on their own core industries and cannot independently complete non-core business logistics activities. One way to solve this problem is logistics outsourcing. Therefore, how to select suitable logistics suppliers has become the key to meet the logistics needs of enterprises. Combining with the existing supplier selection index system and the logistics demand of enterprises, this paper chooses four aspects which can reflect the comprehensive strength of suppliers according to the theoretical basis of selecting third-party logistics suppliers. This paper subdivides these four aspects, summarizes 10 evaluation indicators of supplier selection, and establishes a relatively scientific and reasonable index system. Then, the weight of each index is determined by analytic hierarchy process. For supplier processing, clustering analysis is used to classify supplier types. By comparing and analyzing the advantages and disadvantages of three common-used clustering methods, this paper proposed an improved hierarchical clustering method, which can effectively determine the cluster center coordinates and improve the accuracy of supplier classification. Finally, a comprehensive evaluation method of third-party logistics suppliers is established by combining the weights determined by the analytic hierarchy process with the clustering results.

Petri net (PN) is a classical tool for modeling, simulation, and analysis. With the emergence of distributed manufacturing system (DMS), PN has been evolved into different forms such as colored PN (CPN) and timed PN (TPN). To fulfill the practical requirements, CPN has been extended to a larger and more complex model. Meanwhile, DMS becomes an important issue for industry, and how to model a complex manufacturing network for better throughput needs further investigation. A simple single type of PN only cannot fulfill all contemporary requirements at the same time. This paper proposes a new concept for a new PN-hierarchical timed CPN (HTCPN), which can be a powerful tool for modeling, simulation, and analysis for current large complex distributed manufacturing system. This paper presents a general structure of DMS and uses HTCPN to model the manufacturing system as discrete event with dynamic behavior. The proposed approach shows superiority on modeling large complex manufacturing system as it represents the material flow and transitions clearly and provides overview and detailed description for DMS.