Hesuan Hu

Liveness Enforcing Supervision of Video Streaming Systems Using Nonsequential Petri Nets

Internet-motivated video streaming systems face such complicated issues as a high degree of network-resource sharing amongst many flows, which potentially leads to deadlocks. Using siphons and their corresponding dangerous markings, this work investigates a method to enforce control iteratively. At each iteration, a generalized mutual exclusion constraint is produced to keep only those markings under which liveness is enforced. Furthermore, a generalized elementary siphon control method is proposed such that the final supervisor is structurally simple. Examples are used to illustrate the proposed approach.

Deadlock-Free Control of Automated Manufacturing Systems With Flexible Routes and Assembly Operations Using Petri Nets

In the context of automated manufacturing systems (AMS), Petri nets are widely adopted to solve the modeling, analysis, and control problems. So far, nearly all known approaches to liveness enforcing supervisory control investigate AMS with either flexible routes or assembly operations, whereas little work investigates them with both. In this paper, we propose a novel class of systems, which can well deal with both features so as to facilitate the control of more complex AMS. Using structural analysis, we show that liveness of their Petri net model can be attributed to the absence of undermarked siphons, which is realizable by synthesizing a proper supervisory controller. Moreover, an efficient method is developed and verified via AMS examples.

Low-Cost and High-Performance Supervision in Ratio-Enforced Automated Manufacturing Systems Using Timed Petri Nets

In the context of automated manufacturing, this work proposes a new special class of timed Petri nets, namely, Timed ratio-enforced Augmented Marked Graph (TAMG) and its low-cost and high-performance supervisor synthesis methodology. A supervisor is composed of a set of control places (monitors), each of which is easy to be algebraically specified by a generalized mutual exclusion constraint (GMEC) to prevent certain siphons from being undermarked. In order to make a good tradeoff between the supervisor implementation cost and system performance, a mixed integer programming (MIP) approach is formulated to synthesize the monitors. An example is used to validate the effectiveness and efficiency of the proposed method. The results show that the proposed method remarkably outperforms any existing ones.

Algebraic Synthesis of Timed Supervisor for Automated Manufacturing Systems Using Petri Nets

For practical automated manufacturing systems (AMSs), the time dimension is of great significance and should be integrated in their plant models. Reasonably, many of the realistic general mutual exclusion constraints (GMECs) imposed on these discrete models should be timed rather than merely algebraic or logic. In the past, such a problem was studied on the basis of the Ramadge-Wonham supervisory control technique (SCT) and the theory of regions. It proves to be NP-hard since it necessitates the generation of reachability graphs. This paper shows that it can be solvable in polynomial time by using generalized linear constraints, which are originally proposed to increase the expressive power of the linear marking constraints. By dividing each constraint into marking, firing vector, and Parikh terms, its respective control place can be synthesized algebraically without considering the separation of dangerous states and events. Several examples are used to validate the effectiveness and efficiency of the proposed approach.

Supervisor Design to Enforce Production Ratio and Absence of Deadlock in Automated Manufacturing Systems

This paper proposes a new Petri net class, namely, Ratio-enforced weighted Augmented Marked Graphs (RAMGs), and solves ratio control and liveness-enforcing supervision problems for automated manufacturing systems. RAMGs can ensure the required product ratios as demanded by production planners. Since the deadlock of such a system can be attributed to improper acquisition of finite shared resources, a supervisor is introduced such that they are properly allocated. This paper proves that ratio and supervisory controllers for an RAMG can be separately designed. Their design methods are presented. Examples are given to illustrate them.

Supervisor Optimization for Deadlock Resolution in Automated Manufacturing Systems With Petri Nets

For automated manufacturing systems (AMSs), deadlock resolution in terms of Petri nets remains an attractive topic to which many approaches are dedicated. However, few of them can quantitatively optimize certain indices during their supervisor synthesis process. This causes unnecessary control limitations and often leads to high implementation cost. In the framework of Petri nets, this paper proposes a method to synthesize a cost-effective supervisor with the aid of a set of mathematical programming formulations. Along the same vein, we also show some results by investigating timed Petri nets, which can be utilized to make a good tradeoff between implementation cost and system cycle time. Examples are used to validate the effectiveness of our result.

A Petri Net-Based Discrete-Event Control of Automated Manufacturing Systems With Assembly Operations

In the context of automated manufacturing systems (AMSs), Petri nets are widely adopted to solve the modeling, analysis, and control problems. So far, nearly all known approaches to liveness-enforcing supervisory control study AMSs with flexible routes, whereas little work investigates the ones with synchronization operations. Compared with flexibility, synchronization allows the disassembly and assembly operations that correspond to splitting to and merging from different subprocesses, respectively. Such structures bring difficulties to establish a liveness condition of the Petri net model of AMSs. In this paper, we propose a novel class of systems, which can well deal with these features so as to facilitate the investigation of such complex systems. Using structural analysis, we show that their liveness can be attributed to deadlock freeness, which is much easier to analyze, detect, and control by synthesizing a proper supervisory controller. Furthermore, a set of mathematical formulations is proposed to describe and extract the corresponding deadlocks. This facilitates the synthesis of liveness enforcing supervisors as it avoids the consideration of deadlock-free but nonlive scenarios. The effectiveness and efficiency of this new method is shown by AMS examples.

An Optimization Approach to Improved Petri Net Controller Design for Automated Manufacturing Systems

Sensors and actuators are two indispensable parts in the paradigm of feedback control. Their implementation cost should be properly evaluated and constrained. In the previous work, a Petri net monitor with the least cost is synthesized through integer programming formulation. Despite its technical correctness, the existing method may lead to undesirable results when the net structure contains some shared or unshared resource places of a manufacturing-oriented net model. A necessary and sufficient condition is established to show that certain structures can lead to deadlock-prone supervisors. An efficient algorithm is developed to identify such structures. Furthermore, it is shown that if one can identify such structures at the initial stage, it is possible to achieve desirable controllers for the original systems. The theoretical correctness of the proposed algorithm is discussed. A manufacturing example is provided to illustrate the proposed approach.

Liveness and Ratio-Enforcing Supervision of Automated Manufacturing Systems Using Petri Nets

In automated manufacturing systems (AMSs), Petri nets are widely adopted to solve supervisory control problems. Among them, how to enforce liveness and fairness for AMS constitutes an important problem. Enforcing liveness avoids the occurrence of deadlock situations to inhibit the emergence of partial or complete blockness for processing jobs in AMS. Fairness requires the determination of a reasonable regulation scheme for orderly resolution such that a desired ratio can be assigned among different processes. This paper proposes an efficient method to design supervisors that enforce both, which is based on the invariance property of Petri nets. A novel approach is proposed to iteratively identify empty siphons as solutions to a set of linear inequalities. Supervisors are then designed to control these siphons. The applicability of the proposed methodology is illustrated through examples. Its comparison with existing works is presented to demonstrate its advantages.

Supervisor Simplification for AMS Based on Petri Nets and Inequality Analysis

In the framework of automated manufacturing systems (AMS), Petri nets are widely used to model, analyze, and control them. Resolving deadlocks is of paramount significance because their emergence may likely zero a systems throughput, if not necessarily. Supervisory control technique is the most widely adopted method to resolve them. A control policy can be converted into satisfying a set of inequalities, each of which corresponds to a siphon in a Petri net structure. The number of siphons can be exponential in the worst case, so does the number of inequalities. Taking into account the independent and dependent inequalities, this paper proposes a method to remove all the dependent inequalities, while preserving only the independent ones. This method can significantly reduce the size of a supervisory controller. Examples are presented to illustrate the effectiveness and efficiency of this method.