A Study of the Impact of Production Scheduling Using Enterprise Simulation

Full Citation
Helal, M., Rabelo, L., Jones, A. 2004. A study of the impact of production scheduling using enterprise simulation. IERC 04, The IIE Annual Conference;, May 15-17 Houston, TX

Extended Abstract
The strong tendency toward more integration in businesses and the system thinking approaches are shaping the way manufacturing enterprises are being managed, especially in an environment of market globalization, ever varying customer requirements and increasing competition. As a result, the development of manufacturing simulation systems is directed toward integrated enterprise modeling in two directions of integration: horizontally by integration of manufacturing processes with the entire enterprise logistics chain, and vertically by integration of the decision making processes at the strategic, tactical and operational management levels [1]. To achieve a competitive advantage an enterprise must cope with these trends and changes in the working environment.

In this work, we introduce a hybrid simulation approach that combines discrete-event models with system dynamics models to simulate the entire manufacturing enterprise system. This approach basically covers the strategic, tactical, and operational levels of decision-making in a single simulation of multiple models at different levels and resolutions. System dynamics (SD) is utilized to model the higher levels of decision-making in the enterprise, while discrete-event simulation (DES) models are utilized for the factory and shop floor levels. Studying the interaction between strategic planning and the production activities is the core of this work.

Discrete event simulation has been the most usable simulation tool in manufacturing applications. It has the ability to describe the most complex systems at any level of details while allowing tracking the status of individual entities and resources and estimating numerous performance measures associated with these entities under various operating conditions. This makes it the most appropriate approach to simulate the detailed manufacturing activities.

At the higher levels of decisions making, which are mainly non-manufacturing tasks, the detailed approach of DES is not as appropriate [2,3,4]. For one reason the decision makers at these levels do not prefer detailed analyses. In addition unlike manufacturing tasks data, the non-manufacturing tasks data are only available as rough estimates [5]. For that we consider SD simulation, which is a system thinking approach that focuses on the structure of the system and its dynamic behavior due to taken management policies, for simulating these levels. SD is not a data-driven approach and its data requirements are minimal. The central concept in SD is that all the components of a system interact through causal relationships that come about through feedback loops. Thus a model is a dynamic picture of the perceived cause-and-effect relationship among system components. It can easily incorporate continuous and qualitative systems parameters as well. These features make SD well appropriate for simulating strategic and tactical decision making processes.

It is understandable that the driving force for a company’ strategic planning is its need and desire to manufacture competitive products. Operational scheduling on the other hand refers to the short-term decision problems in executing the company’s business strategies and achieving the planned goals. Traditionally planning and scheduling used to be very distinct phases. Such distinction does not match the requirements of the current operating environments. However a practical holistic approach to coordinate and integrate them is not yet available. The proposed hybrid simulation approach offers this holistic approach.

We consider a semiconductor enterprise, which typically works in a very dynamic and capital intensive industry. Top management decides based on market analysis to introduce a new generation of chips or enter a new electronics market. Scheduling receives a set of orders and designed goals with predefined release dates for the raw materials and due dates for finished products. Given the time consuming processes involved in semiconductor manufacturing operations, scheduling is a complicated task that would involve the design of new processes that should consider the cost constraints of the management. The related strategic resource allocation decisions made at the enterprise levels in order to be implemented at the production level have strong impact on the effectiveness of the scheduling and other production functions. Reciprocally the effectiveness in executing top management plans to supply the required demand and in using the allocated resources at the operational and tactical levels has direct impact on the ability of the enterprise to satisfy its market commitments and on its financial performance as well.

Further when more than production plants exist within the enterprise (which is the case for a wide range of current manufacturing enterprises) the situation becomes more complicated especially when the different plants work at different levels of performance and have different impacts and contributions to the enterprise’s profitability.

Planning, marketing and customer relations management functions at the strategic levels must be in close coordination with the shop floor level functions. Real time data reflecting the shop floor status must be available during the review of the business strategies and during deciding on introducing new products or entering new markets as well as during making the resources allocation decisions. These decisions must be made fast enough to seize available market opportunities which are always the target of many competitors. In addition to being fast, they should be reliable and based on feedback from all participants, so that later modifications and corrections, which are usually more costly, are eliminated or minimized.

This proposed hybrid simulation environment provides the practical framework to achieve the needed integration. It is utilized to study the impact of the top management decisions (such as resource allocation and new product decisions) on developing production schedules and capacity management at the operational levels. Equivalently it allows for studying and monitoring the execution of strategic plans at all levels. The backbone of the methodology is the use of the feedback information going back and forth between the SD model (higher management levels) and the DES models (lower management levels). Management policies can then be tested in an overall approach, before being executed. This also leads to better communications among all levels of management and can increase the role of the tactical and operational levels in the design of the strategic policies.

Keywords: Scheduling, Enterprise simulation, System dynamics, Hybrid simulation