Queuing Theory for Semiconductor Manufacturing Systems - Article Example

The paper mainly entails application of the queuing theory within the semiconductor. Nevertheless, traditional queuing theory can never be used within the semiconductor manufacturing systems since it has numerous disadvantages in its application. This is due to the failure of the traditional theory in the reduction of the cycle time. The present queuing models also focus mainly on the underlying simple extensions of the prevailing classical queuing theory, which does not work in the complex semiconductor manufacturing systems. Critique The abstract of the article is mainly to attract and enlighten the prevailing readers in comprehending the content of the paper. The paper is based on the queuing theory and its corresponding applications within the manufacturing industry in regard to the semiconductor manufacturing systems. In a nutshell, the abstract of the article explains the purpose whole paper. The introduction part of the article is fascinating to certain degree since the authors elucidate the prevailing steady pressure applied on the semiconductor manufacture in order to reduce the cycle time and improving delivery performance. Nevertheless, the process of application of steady is normally cumbersome task for manufacturers due to problems from sophisticated tools and corresponding specified procedure. The prevailing authors mainly elucidate particular problems that ought to be solved in service processes in regard to toolsets, which are highly variable. Service processes flow possesses deviations and involves simulations, which is one of the approaches for approximation of the cycle time. Moreover, the section is normally better when bird eye’s perception is applied in gauging the impact of alteration of available capacity, examination of the underlying effect within the quantity and product mix. This correspondingly grants action plan for advancing the operational planning and fabrication performance. Queuing Models for the Tool Sets part is associated to specific models of semiconductor manufacturing system, which is categorized by the authors as single machine stations and corresponding multi machine stations. The authors explain the models and systems to aids the reader in comprehending the systems thus depicting the level of appreciation of the content of the article by the authors. Moreover, the authors elaborate the underlying weaknesses of the systems by displaying the precise analytical solution that is extremely cumbersome to achieve via application of single machine stations. Simple machine station is not normally utilized in describing accurately the fabrication performance since they are very use over simplified models. The reader is therefore advice to offer some probable solution for advancement of the performance of simple model after the authors have explained association amidst the capacity and the work processes of the simple machines. The authors strive to make the audiences to comprehend multi machine stations by offering formula employed in the computation of performance of the machines in section II and corresponding sub section B. The formulas are mainly based on the assumptions and the answers are prone to be wrong in case one of the assumptions is varied. For instance the formula is based on the performance of diverse machines or corresponding diverse sections of the machines performance. The authors ought to discuss assumptions within their prevailing article in case the machine fails since the performance of the whole machines is the determinant of the precise measurement of the performance. According to the authors in regard to the sub section batch tool, thirty three percent of semiconductor manufacturing systems utilize batch processing. These kinds of batch services are required in other field of research. Therefore, the authors ought to elaborate on particular research works to be undertaken in order to provide adequate information of the systems. Within the subsection of the multiproduct and corresponding setup, the author talks about the lofty product mix that has numerous problems in regard to it multiproduct queuing systems, controlling policies ,scheduling dispatching, and switching amidst the tool and the products. The authors suggest particular fundamental estimations to counter the problems, setting up of minimization and avoidance and type priority. Brief explanation of the underlying four steps mainly aids the readers in boosting the understanding of the rules. The authors propose the prevailing historical data that are employed in forecasting of prospect time within the subsection in regard to empirical estimation. They undertake this by recommending certain minor adjustment that are essential for making realistic approach since the alteration of approximation normally eliminate overestimation that are founded on the experimental interpretations. Moreover, the authors also recommend same approach to be applied to computation of the ration amidst the actual performance and corresponding approximation model. Nevertheless, the authors ought to elaborate on the tool and techniques since the details of the explanations are inadequate in regard to the subject and the significance of the topic. Third section mainly talks about queuing networks and the authors stipulate that semiconductor manufacturing systems is dependent on the rival process of the corresponding tool test. Queuing network of the system can be molded by making flows arc and toolsets nodes in the manufacturing process. I perceive this recommendation to be extremely substantial since it is applicable within the SMS thus making the underlying system an open network. Moreover, the section of the report is considerably inclusive since it offers a genuine approach. The decomposition approach section of the article grants new approach in regard to decomposition thus making it more applicable within the manufacturing and it is currently known as the queuing network analyzer. Nevertheless, it mainly depends on certain assumptions thus making the system more complex when applied hence making it cumbersome in all environment. The nest subsection mainly talks about the fluid network and the authors elucidate that queuing network can normally be reshaped or altered into the fluid network by assuming that every process. Nevertheless, the problem comes in when there is either failure or stoppage within one of the process which normally makes other processes to idle. This makes the performance is immeasurable as the fluid network, which is purely a linear network of the inflow and corresponding outflow. Network cannot be applied for the approximation of the underlying cycle time that is extremely essential for assessing the entire performance. Moreover, network will be fundamental in case network drains in regard to the finite time commencing with the buffer level. The subsection D of the article talks about the diffusion estimation for the congested network. Congested network is mainly categorized into close and open network. Diffusing estimation can be utilized for studying specific scheduling policies according to the authors. This mainly entails first and second moments that possess some disadvantages that pertain to controlling policy for the lofty urgency products, which is non-recoverable in regard to this method. The system mainly relies on numerous assumptions than any other systems thus the authors ought to have highlighted such drawbacks within the article. The writers comments of the fundamentals problems in regard to the systems that control processes in subsection E. Each system possesses particular characteristics. Nevertheless, particular shortcomings such assessment of performance because the underlying non-control actors or deferral within the process. The writers recommend Lyapunov quadratic functions for studying of the performance of the steadiness of queuing network foundations systems. Literature review of the paper is gives limited and insufficient outcomes concerning the semiconductor manufacturing systems and corresponding manufacture practices. The authors propose that more examination can be undertaken to advance queuing modeling quality thus demanding more research in future by the researchers. Recommendations There is numerous other models that can be utilized in order to make this paper more comprehensive and reliable. Other models that can be employed are general models that are mainly batch processing model and encompass front opening unified pods,multi-class open queuing network, multiple server coupled with chamber tools, models operation based on the preemptive interruption and non- preemptive interruptions model. These models can be incorporated within the literature review and finding to make the paper more precise. Additional model that can be utilized is the model for enterprise production planning and optimization that are being utilized by the IBM. EOPS is taken as the best model because of the complexity of the semiconductor manufacturing system. Moreover, queuing model of the EPOS is feasible and possess fundamental features of the semiconductor manufacturing system, which can advance the outcomes and the corresponding accuracy of the analysis of the queuing within the environment of the semiconductor. I therefore recommend the authors to take into consideration the above models in order to offer full solutions for the problems. Work Cited Mönch, Lars, J W. Fowler, and Scott J. Mason. Production Planning and Control for Semiconductor Wafer Fabrication Facilities: Modeling, Analysis, and Systems. New York: Springer, 2013. Print Read More