روش پایگاه داده رابطه ای برای برنامه ریزی مبتنی بر سیستم پشتیبانی تصمیم گیری خطی برای برنامه ریزی تولید ثانویه تولید محصول چوب

Secondary manufacturers in the forest products industry face a complex production planning process. Linear programming (LP)-based applications have addressed this production planning issue. However, most models have been developed for a specific plant configuration and cannot readily be applied to others. A relational database approach was used to create an integrated linear programming-based decision support system that can be used to analyze production planning issues in a wide variety of secondary wood product manufacturers. The flexibility of the resultant system indicated the potential to analyze production strategies in the highly dynamic environment characteristic of secondary manufacturers.

The expansion of the secondary manufacturing industry is a common interest for the major timber producing, importing and exporting regions of the world [19]. Governments recognize the potential of the secondary industry to stimulate economic development through job creation and economic diversification in forest dependent economies [6], [20] and [21]. Furthermore, rationalization in the primary manufacturing sector has resulted in fewer but larger more efficient sawmills in the ECE region (Europe, North America and the CIS) [19]. This has resulted in downward pressure on employment in the primary manufacturing industry and may exacerbate the need for increased activity in the secondary industry to support these economies. Secondary manufacturers are typified by a wide variety of raw materials, manufacturing processes and potential products. Consequently, the production planning process can be highly complex. This complexity is further increased by the dynamic nature of the market environment [16]. A limited number of linear programming (LP)-based applications have addressed this complex production planning issue in secondary manufacturing. However, most models have been developed for a specific application domain and cannot readily be applied to others [2], [3], [5], [12] and [15]. This may help explain an observation made by Carino and Willis [4] regarding the lack of optimization modeling applications in small and medium forest products companies. Srinivasan and Sundaram [18], and Muhanna and Pick [11] convey the importance of flexible approaches to model design, in order to broaden the scope of potential applications. Furthermore, Olhager and Rapp [14] suggest that specific operations research applications may have a short life span due to the highly dynamic nature of the manufacturing environment. The research objective was thus to develop a generic linear programming-based modeling application that will provide a flexible DSS for production planning in a wide variety of secondary manufacturing plants. The DSS was implemented and validated in a secondary manufacturing operation in British Columbia, Canada and is currently being used for production planning purposes at this operation. This paper presents the architecture of the resultant DSS and the mathematical formulation to the underlying LP. Examples of generic operations research models that incorporate a high degree of application flexibility can be found in a variety of manufacturing environments. Metaxiotis et al. [10] present an object oriented approach to an advanced DSS that is integrated with a Management Information System. The software package was integrated with a management information system in a customized industrial environment in Greece, and uses dynamic simulation techniques to provide production planning and scheduling on a daily basis. Fourer [7] implemented a generic linear programming model via a database system for production planning in the American steel industry. The design of the database is closely related to the structure of the mathematical formulation. Gazmuri and Arrate [8] discuss the development of a system to build optimization models for a variety of applications. They employed a number of software tools from a prototype of a modeling system, and implement a general production-planning model in an appliance manufacturer in Chile. Zhang [22] provides the only example found by the authors in the secondary manufacturing sector of the forest products industry by developing a multi-period model for furniture manufacturing. Flexibility is provided by viewing the manufacturing operation as a number of processing stages that possess generic attributes, rather than focusing on the specific attributes of each. The model incorporates a user interface programmed in Fortran that generates the linear programming model from user inputs and creates an optimal solution report. The work presented in this research is similar to that carried out by Zhang in that the manufacturing process is considered in general terms. However, a relational database approach is employed to utilize the data management and manipulation capability offered by modern database systems [9] and [13].

Production planning in the secondary manufacturing sector is a complex task. The interactions of raw materials, with process and marketing factors produce complex interrelationships that require simultaneous analysis in order to find the optimal operating strategy. An examination of the literature on operations research techniques for production planning in secondary manufacturing revealed the specificity of existing applications. Most models have been designed for specific secondary manufacturing plants or focus on solving specific problems. The decision support system developed in this research has the flexibility to model a wide variety of secondary manufacturing plants and can be used to address a range of short-term production planning issues. The widely available relational database management tool MS Access provided an appropriate design environment for building the DSS. The relational database provides an efficient data management tool for handling potentially large data sets. Model data can be easily updated through the appropriate forms and re-use of common data is facilitated by their storage in the database. Model-data independence is provided by a database that manages all model data, and SQL queries that define the mathematical model. As noted by other authors [1], [9] and [11] model-data independence facilitates data manipulation, reuse of the same mathematical model in different specific modeling applications, and provides greater design robustness and clarity. The DSS has a broad application domain, and can be used to provide insight on decisions concerned with product mix, raw material sourcing, production strategies, pricing strategies, resource valuation and capital appropriations requests. Furthermore, the life span of the resultant DSS should be increased through the flexible design that allows one to model a wide variety of machine centers and the ability to customize the process flow to model specific production planning scenarios. This flexibility is a major advantage in a manufacturing sector characterized by a highly dynamic operating environment.