IDEF method-based simulation model design and development framework, The purpose of this study is to provide an IDEF method-based integrated framework for a business process simulation model to reduce the model development time by increasing the communication and knowledge reusability during a simulation project. In this framework, simulation requirements are collected by a function modeling method (IDEF0) and a process modeling method (IDEF3). Based on these requirements, a common data model is constructed using the IDEF1X method. From this reusable data model, multiple simulation models are automatically generated using a database-driven simulation model development approach. The framework is claimed to help both requirement collection and experimentation phases during a simulation project by improving system knowledge, model reusability, and maintainability through the systematic use of three descriptive IDEF methods and the features of the relational database technologies. A complex semiconductor fabrication case study was used as a testbed to evaluate and illustrate the concepts and the framework. Two different simulation software products were used to develop and control the semiconductor model from the same knowledge base. The case study empirically showed that this framework could help improve the simulation project processes by using IDEF-based descriptive models and the relational database technology. Authors also concluded that this framework could be easily applied to other analytical model generation by separating the logic from the data.
Simulation is one of the most widely used decision aid tools due to its power, flexibility, and robustness. Particularly the discrete event simulation (DES) can model and analyze the behavior of many real life processes such as business processes, supply chain, and manufacturing processes. However, as Ryan et al. pointed out (2006), the simulation modeling often becomes a heavy programming task with the essence of the system being modeled lost in the detailed programming codes. In this way, the essence of the system is visible only to the code developers. This could create several potential problems for those who are involved in a simulation project. For example, it may create a serious information reusability problem. A simulation model is an abstracted representation of a real system to solve specific problems. Hence the information collected and extracted from the real system should be systematically represented and stored for future reuse in the form of systematic descriptions and formats. It may also cause a communication problem between developers and users. Typically users are domain experts who want to experiment with the simulation model to solve domain specific problems. This task requires frequent parameter changes and modification of the model. However, the heavy codes add difficulty to the proper management of this task. If we consider a simulation model development as a project, and if we have a structured systematic tool to support the simulation project, we believe that these problems could be managed. Sheppard (1983) proposed a widely cited “40-40-20” simulation model development time rule which states that analyst’s time should be distributed as follows for a successful simulation project: (1) 40% to requirement collection phase such as problem formulation, project planning, conceptual model development, and data collection; (2) 20% to model translation phase; (3) 40% to experimentation phase such as model verification, validation, implementation, and interpretation. Hence, for successful implementation of any simulation project, it is particularly important to have a right approach to the requirement collection and the experimentation phases. Hence, this paper intends to provide an integrated framework for those two phases in a simulation project.
The process description methods could play an important role in the simulation requirement collection phase. Although many process design, analysis and modeling (DAM) methods have been developed, using these methods in isolation – non-methodological approach – often fails to capture critical system behaviors due to the complexities and component interactions within the system. A methodological approach – systematic usage of a suite of methods – has a greater chance of success at representing critical system behaviors since it can account for diverse aspects of DAM activities such as information, function, and process interactions by a systematic and integrated usage of methods. IDEF (Integrated DEFinition) is a suite of descriptive modeling methods within which several different modeling languages are defined to describe systems from different perspectives. First, since IDEF is a well defined suite, it is considered to be easier to implement a methodological approach with the IDEF suite rather than with a completely different set of methods. Second since it is a descriptive modeling method, it could easily abstract and capture the essence of the system. In a typical simulation project, a project team consists of many team members such as system analysts, developers and domain experts. The system analysts collect and refine requirements with assistance from domain experts. This is an iterative communication process among all members. The ‘descriptiveness’ of IDEF methods could make this communication process easier and smoother than any other non-descriptive methods. For these reasons, IDEF methods have been a continued research subject.
The first category of the IDEF method related research attempted to build a generic and conceptual descriptive model using IDEF suites in a specific domain (Ang et al., 1994; Zhang et al., 1996). Another category proposed a way to generate an analytical model from a specific IDEF model. For example, an IDEF3 method has been used to generate simulation models using Witness simulation software (KBSI, 1995) and using Arena software (Resenburg et al., 1995). Jeong et al. (2008) developed a scheme to integrate the IDEF3 with a general open queuing network where IDEF3 works as a knowledge repository. The third category employed multiple IDEF methods and attempts to reuse common system knowledge among the different IDEF methods. For example, Lingzhi et al. (1996) proposed a scheme to integrate IDEF1 with IDEF0 for a computer integrated manufacturing information system design. Chen et al. (2004) also proposed a scheme to develop the enhanced IDEF1 information model based on the IDEF0-based process information, which could serve as a base representation for an information model. This paper covers both the second and the third category together. It is an extension of Cho et al. (1999), KBSI (1995), and Chen et al. (2004) in that it attempts to provide an integrated framework of IDEF method-based simulation model design and development to help a successful simulation project.