In this paper, an unified approach is proposed to solve forbidden state problems in a general class of controlled Petri nets (CPNs). A new subnet called the step-aligned subnet, which can describe various kinds of PN structures such as marked graphs, state machines and asymmetric nets, is introduced. The step-evaluation method is developed to efficiently investigate the reachable marking problem in step-aligned subnets instead of the full search of reachability graph. The crucial idea of the method is to decompose a large reachable marking problem into a number of smaller subproblems. This paper derives a necessary and sufficient condition in order to find maximally permissive controls in the context of step-aligned subnets. Theoretic results are applied to modeling and control of a flexible assembly cell.

We concentrate on some important timing characteristics of elevator group control (EGC) systems. The main discussion and calculations should be considered as a further development of the three passages concept. To achieve better global performance of the EGC system, especially to reduce hall call waiting time, a calculation methodology of car kinematics is presented. The importance of detailed kinematics considerations is demonstrated with data taken from a real high-speed EGC system in 63-floor building in Seoul, Korea.

In this paper, a new approach is proposed to control elevator group systems. Hall call waiting times (HWTs) are estimated by a statistic approximation of traffic data, which are reasonably accurate under different kinds of real traffic situations. Two assignment policies are proposed to satisfy the passengers' demands and to handle exceptional situations. In addition, these policies are extended in order to reduce the variance of HWTs and long HWTs. The performance of the elevator system employing the proposed policies are compared with each others in terms of several performance indices by computer simulation.

An optimal group control for elevator systems is proposed with a timed Petri net based approach. The elevator system is modeled by timed Petri nets and hall call waiting times are estimated with a moment generating function method, which is applicable to real traffic patterns. Two assignment policies are proposed to satisfy the demands of passengers (i.e. hall/car calls) and to handle exceptional situations. The performances of the elevator system employing the proposed policies are compared with each others by several performance indices.

A method is proposed for solving the forbidden state problem in a general class of non-ordinary controlled Petri nets (NCPNs). A concept of step-aligned subnets is introduced to handle the behavior of reachable markings in NCPNs. In addition, the step-alignment algorithm is developed to transform any loop-free subnets into step-aligned subnets. New procedures called the step-evaluation method and the forbidden condition reduction algorithm (FCRA) are developed to efficiently check admissible/boundary conditions given forbidden conditions and to reduce the computational complexity.