Prediction Model Using Coverability Tree from a Modeling in Petri Nets Applied in AGVs Dispatching

Abstract

In the last years, the industries are applying automation techniques with the aim to increase their efficiency to remain competitive. Due to this fact, there is an intensive search for techniques and methods applied to manufacturing systems for improvements, whether in quality, service deadlines and / or increased production. Several studies and practical applications have indicated that the Automated Guided Vehicles (AGVs) are efficient for transport task in industries and warehouses. The management of these AGVs is the key to a transportation system that ensures the improvements envisioned by industries. One of the main problems encountered in the management of AGVs is the decision to dispatch. Some authors suggest that a weak point of dispatching rules, even the multi-attributes, is to consider only the values of variables in current time for decision-making. This paper proposes a prediction model in which one achieves an improvement in the optimization objective by reading from specific states of the factory in the near future. The proposal is based on the use of coverability tree from modeling in Petri nets with ability to provide important data for AGVs dispatch system. From these data, the dispatch system can take more assertive decisions to optimize the performance of Flexible Manufacturing Systems (FMS). The tests are performed using the software CPNtools, to model Petri nets, and Simio software to build the virtual environment, thereby having two different levels of abstraction. The validation is done through the analysis of scenarios that can happen in a production system, and with the use of the proposal of prediction is verified that in these scenarios is possible to extract important information.

Flow-Shop Predictive Modeling for Multi-Automated Guided Vehicles Scheduling in Smart Spinning Cyber–Physical Production Systems

“…Concerning the limitation on processing equipment, a machine processes only one raw material at a time, and this constraint is determined by the attributes of the processing equipment for different processes in the drawing shop, as shown in Equations ( 8) and (9). When p = 1, the raw materials for machining may come from different production lines.…”

“…Though the benefits of allowing for AGV job preemptions can be significant, few studies have considered pre-emptive scheduling in the context of flow-shop predictive modeling. For example, in [9,10], AGV dispatching and cooperative waterborne AGVs prediction models were addressed, and a class of precedence constraints was proposed.…”