Search

In the modern market, keeping high competition in brands and varieties in type of products is the way for survival of manufacturing industries. Therefore production control methods with capabilities of quick responses to rapid changes in the demand and efficient distribution of the raw material throughout the network are of importance among leading manufacturers. Nowadays, the production control problem has been widely studied and a lot of valuable approaches including queuing theory, Petri nets, dynamic programming, linear programming, hybrid systems were proposed and some of them are implemented. Though up to this moment many methods have been developed, the factory performance remains a challenging problem for further research. Motivated by this problem we study the performance of several manufacturing networks operated by surplus-based control. In the surplus-based control, decisions are made based on the demand tracking error, which is the difference between the cumulative demand and the cumulative output of the network. The studied networks are a single machine, a manufacturing line, a multi-product manufacturing line, a re-entrant machine and a re-entrant line. The performance analysis is based on the performance factors such as demand tracking errors and inventory levels. Specifically, given the presence of unknown but bounded production speed perturbations as well as demand rate fluctuations, we investigate how close the cumulative production output of a manufacturing network follows its cumulative production demand under a surplus-based control policy. The research is subdivided into theoretical analysis, simulation-based analysis and experimental analysis parts. Theoretical analysis By means of analytical tools, the relation between the production demand tracking accuracy and the inventory levels of the networks is investigated. In order to find this relation, classical tools from control theory are used. Models of production flow processes are formulated by means of difference as well as differential equations. In order to analyze their performance, optimal control theory and Lyapunov theory approaches are exploited. Simulation based analysis By means of simulation tools, the theoretical results on performance are evaluated by time-based simulation models. Thus, all theoretical results are illustrated and confirmed by computer simulation. Also two comparative studies are conducted. The first comparative study is realized in order to test the theoretical results on more accurate models, which are event-based. The results are shown to be in agreement with the theory. The second comparative study is on time-based models, where the behavior of a line, a single re-entrant machine and a re-entrant line is tested under three commonly used surplus-based production policies. The performance of each network is evaluated and the results are presented. Experimental analysis An experimental prototype is invented, designed and developed for education and research purposes. The prototype is a hardware tool that serves as a liquid-based emulator of manufacturing network processes. In its core, the liquid-based emulator consists of several electrical pumps and liquid reservoirs. The electrical pumps emulate manufacturing machine behavior, while the liquid reservoirs serve as the intermediate product storages, also called buffers. In the platform, pumps and tanks can be interconnected in a flexible manner. In that way the prototype permits an easy and intuitive way of studying manufacturing control techniques and performance of several network topologies. A detailed system description is provided. Several network configurations and experimentations are presented and discussed.