Simulation in Supply Chains: An Arena basis

The quest for global competitiveness brought about new business approaches, of which the supply chain has become an important entity during the last few years. With even more complex decision structures, demand variation and the need for evaluating alternatives within this frame, simulation and simulation-optimization have been identified as key decision-making tools. This paper briefly reviews the basic characteristics of supply chains, and illustrates that existing software may be integrated towards a supply chain simulator. Opsomming Die strewe na globale mededingendheid vereis nuwe benaderings deur ondernemings, terwyl die toevoerketting 'n belangrike entiteit gedurende die afgelope paar jaar geword het. Toenemende kompleksiteit in besluitneming, variasie in vraag en die behoefte om alternatiewe binne hierdie komplekse raamwerk te evalueer, het tot gevolg dat simulasie en simulasie-optimering as sleutel-besluitneming gereedskap beskou word. Hierdie artikel gee 'n kort oorsig oor die basiese eienskappe van toevoerkettings, en dit word getoon dat bestaandeprogrammatuur integreer kan word om 'n toevoerketting-simuleerderte ontwikkel. http://sajie.journals.ac.za


Introduction
Globalisation contributed to the evolution of informed, demanding customers, which requires a new operational approach to satisfy them.This operational approach evolved into the supply chain and its management, and today the supply chain is a given in many enterprises.This brought about new management complexities that require understanding and analysis of the supply chain in order to stay competitive.
The modern customer profile includes the demand for quality products supplied in any quantity within a short time.Time , quality and cost are therefore basic, but important measures (Kaplan[1 D. Time is a key weapon in today's competition, while quality has been a critical competitive dimension since the 1980s, and is still important today.Customers are always evaluating the price they pay for a product, and usually measure the complete package (lead time, product quality and service quality) against the price paid .
Enterprises are required to collaborate in both the design and production processes, while information must be shared.It is also required that coordination takes place on the strategic , tactical and operational planning levels, which impacts the business processes and the information system.While supply processes are becoming more and more important, their complexity and requirements pertaining to quick response times increase .An optimal design of the relations between deliverer and consumer is of essential importance, because today the time between ordering and availability may already be a matter ofhours.Given these complexities and wide range of entities that must be integrated, simulation can be a powerful tool to analyse and eventually optimise the complete supply chain.The supply chain is dynamic due to variation of customer demand, as well as characteristics of its subcomponents.As short time-tomarket has to be achieved without a decrease in income , it is essential to evaluate new products or new operational processes before production is started.This paper outlines some of the basic characteristics.ofsupply chain management, as well as the relationships among them .A supply chain simulator is proposed, using MS-Access and MS-Excel ' files as data sources and targets for simulation reports , while the simulation kernel is implemented in Arena'.The simulator enables users to observe and analyse the dynamic behaviour of a supply chain with this new configuration.

Basic characteristics of supply chain management
Some of the basic characteristics of supply chain management are identified in this section.

Definition ofsupply chain management
Supply chain management (SCM) is defined by Umeda e.a.[2] as ''the management of material and information flows both in and between facilities in the chain, such as vendors, manufacturing plants, and distribution centres".
I Excel and Access are registered trademarks of Microsoft Corporation .
2 Arena is a registered trademark of Systems Modeling Corporation.

Time horizons
Management and planning occur on three levels, while these three levels each has specific characteristics, as shown in figure I (Umeda [2], Ingalls [3]): ~trategic planning ~Longest time horizon (up to several years).

Supply chain management processes
When considering the processes involved in supply chain management, it is useful to refer to the Supply Chain Operations Reference-model (SCOR) (SCC [4], Weyers [5]).This model describes a supply chain in terms of four modelling constructs or processes, which are : • Plan • Make • Source • Deliver Using these four processes, the supply chain is modelled and described as shown in figure 2: Figure 2 The SCOR Management Processes It is evident from the previous figure that various enterprises can be integrated into the supply chain, with at least one interface between enterprises.The scope of these processes is briefly as follows (SeC [4]): Plan (Demand/Supply planning) Some characteristics ofthe supply chain were outlined in the previous section and it should be clear that managing the supply chain is difficult due to its complexity.Industrial engineers are equipped to assess a system like a supply chain in its entirety.and they nearly always apply quantitative methods for improvement of and better decision making in a system.These methods include linear programming, integer/mixed-integer programming and network models (Hicks [6]).Supply chain issues are mainly of strategic and business planning nature.and problems are of diverse types such as: • Inbound-materials management: Location, acquisition and preposition materials.
• Finished-goods inventory management: Inventory levels, Inventory location(s), Make to stock or make to order.• Supply chain strategy issues: Make or source manufacturing process supplies.
The problems in supply chain management involve systems dynamics, while variation must be assumed as a given .Quantitative methods often fail to satisfactorily address complex problems where randomness is present, and in such cases simulation modelling and analysis is considered.A new dimension is, however, also part ofthe supply chain problem solution requirement, namely simulation-optimisation.This technique combines the capabilities of simulation (modelling complex systems with variance) while utilizing the strengths of various techniques (e.g.evolutionary algorithms) to evaluate different alternatives to fmd an optimal or near-optimal solution.The use of simulation in supply chain management will be discussed in the following section.

Supply chain management and Simulation
Supply chain management is required to develop the business activities for suppliers and customers, and to increase their benefits.Supply chain simulation is useful to speed up the integration of a supply chain, because it allows the evaluation of "what-if' scenarios, with various operational parameters.
In this section a simulator that is able to evaluate some scenarios is discussed.An overall view ofthe possible use of a virtual supply chain manager is also discussed.
The main elements ofthe simulator are: • a simulation model (Arena model) which is the kernel of the system • a Microsoft Access database (MS-Access) with the customer orders (which can be updated on account ofthe prior months order tendency , the prior year order level etc.) • Microsoft Excel worksheets (MS-Excel) for input data (different simulation parameters) and output data (simulation report) .
Although dedicated supply chain software exists (IBM's Supply Chain Simulator (Bagchi e.a.[7]), Promodel Corp.'s Supply Chain Optimisation Suite (Hicks [8]), it is illustrated here that simulation with an existing package like Arena is possible, while integrating it with a standard relational database system as well as a spreadsheet application.This is possible because Arena supports automation and is able to exchange data with Microsoft applications.
The subsequent paragraphs give details about a supply chain simulator that was developed by the authors .

The supply chain under study
The supply chain has six enterprises with two components suppliers (CSI and CS2), one transformation factory (F), one assembly fjlctory (AL) and one warehouse (W).CSl, CS2, F and the AL enterprises have their own carriers, respectively named Cl, C2, C3 and C4.It is assumed that customers go to the warehouse by themselves, so there is no carrier for the warehouse.

The supply chain simulator
The simulator's architecture is based on three existing software applications.These applications exchange data and most of these exchanges are driven by Arena Visual Basic for Applications (VBA) procedures, while only one Microsoft Excel procedure is required to import some Access data (customer orders, order fulfilment lead time) when the simulation terminates.Figure 4 shows the conceptual exchange architecture between the applications through VBA for Arena and for MS-Excel.orner me EXCEL

Figure 4 Conceptual application integration
A supplier manager is implemented in the Arena application.It is a dynamic element for the integration of these applications in simulating the supply chain.It is aimed at coordinating the supply chain activities , by informing other enterprises in the supply chain of a change in customer demand.Thus, these enterprises can react according to their stock levels and the increase in order level.They: • order enough raw materials for the production, • increase their finished goods stock level, and • may deliver these products before the dispatch quantities are completely made up.
MS-Excel provides the user interface to the management system.Thus only a few MS-Excel worksheets are accessed from the user's point ofview.These contain simulation parameters, a report and a few Active:x® controls (command buttons).One of the sheets is used to run a process that: • saves the MS-Excel data, • opens the Arena model and runs it, and • closes Arena and writes the simulation report in MS-Excel.
The kernel of this simulator, namely the Arena model, will now be discussed.

The "Factory" module
There are three role players in this module: • the factory itself, with product and order flows, • the carrier whose trucks deliver the fmished products (product flow) , and • the initialisation section that fills the Finished Products stock when the simulation begins, with a level the user assigns.
Raw material arrives from CS 1 to the factory (its SOURCE), where it waits for finished product replenishment orders.When an order is placed, the raw material entities are released to the factory's MAKE section where the raw material is processed.This process is realised via a global delay, but it is possible to design a complex model for this part.The process time is one of the parameters that can be altered in the MS-Excel worksheet.When the manufacturing process is completed, finished product entities are admitted to the finished product inventory.These products are thus buffered to wait for finished product orders from the Assembly line.When such an order is placed, the entities are released to the factory's DELIVER part.
The stock levels within the factory are also controlled: When the finished product level is lower than a minimum level assigned by the user, an order for finished product replenishment is placed.The order quantity ofproducts is limited by a maximum level that the user has assigned.The finished product level can therefore not be higher than this maximum level.
Finished products are made available to the carrier and wait until: • the carrier truck is full, or • the carrier truck load reached a level the user assigned, or • the last finished product loaded in the carrier truck is the last product in the order.
These decisions depend on variables that the user can assign before the simulation, and that the supplier manager can change when the model runs.
All the modules (CSI, CS2, Factory, Warehouse etc.) have exactly the same properties as the Factory module explained above.Differences come from the fact that some of them have none, one or two raw materials entrance point(s).
The "Replenishment" module The "Customer" module and the Supplier Manager The module Customer is aimed at writing and reading data to and from MS-Access in order to create the customer orders, and to send them to the Warehouse module.The customer orders wait until the simulation time is equal to the user assigned lead time.The implementation ofthis module is included as an example, and is shown in figure 5: The Supplier Manager is implemented in this module because it supports the upstream suppliers with the variation in customer demand.The Supplier Manager is aware of the quantities of the customer orders.Ifthis quantity is higher than the average quantity, replenishment signals are sent upstream for fmished products, and if needed, also for raw materials and assembly.

The Information support components
The se components are implemented in MS-Access and MS-Excel.

The Microsoft Access component
This component contains business transaction information, for example data on orders , which may contain an order number, quantity required and the time the order is due.An example ofa transaction table is shown in figure 6

Figure 6 Typical Order transaction table
The simulation kern el reads data from this component when orders are received.http://sajie.journals.ac.za

The Microsoft Excel component
This component defines the user interface ofthe simulator, which includes a parameter specification area, as well as a reporting area.The latter is filled after a scenario is simulated.A typical interface is shown in figure 7, while a report is shown in figure 8:

Evaluation and Analysis
Some aspects of the behaviour of the supply chain is illustrated in this section.Three specific aspects are: • the effects of a large increase in customer demand., • changing the global proce ss time in the transformation factory, and • fluctuating customer demand.
Many other aspects may be investigated.,but it is sufficient for illustration to discuss only a few simplifications here.The process time of the Factory as well as the time between order arrivals were made deterministic for the purpose ofthis illustration.

Large increase in customer demand
A group of20 orders equally spaced over time is placed on the Factory, each demanding 60 units ofthe product.At a point in time there is a sudden increase in customer demand -500 units are required instead of the usual 60.
The simulator was allowed to handle this situation with and without the Supplier Manager.The results are shown in figure 9: For the given processing time, the supply chain recovers quicker with lower lead times when the Supplier Manager is active in the simulator.Note that the order data was imported from MS-Access, while the graph and its underlying data were created by MS-Excel after the data was exported by Arena.
Increase in process time of the transformation factory The same situation described above can be re-evaluated with a longer processing time of the Factory.Suppose a different enterprise is introduced into the supply chain, with an overall processing time of 0.85 time units.This enterprise may be considered due to its lower product unit cost.Ifthe same demand pattern is fed to the simulator, it yields the result as shown in figure 10: The longer processing time causes a longer recovery period, with longer lead times associated.The Supplier Manager still improves the lead time and recovery period.This example also confirms the complexity of decisionmaking in the supply chain.Although the lead times may be longer, this factory with the longer processing may be a feasible option if overall costs are considered.This would include the lower product unit cost of this factory, the possible higher inventory levels on the fmished product, as well as raw materials upstream.

Fluctuating customer demand
The previous two examples considered the effects ofa single upward increase of customer demand.The following example (figure 11) shows the effects of cyclic deterministic demand, for the original processing time of 0.25 time units: The supply chain can again recover sooner when the Supply Manager is active, given that the resources are not overloaded.
Only specific aspects of the simulator were shown in the examples above.It is important to realise that the supply chain should be evaluated as a whole, and not by isolating some subcomponents of it.However, a subcomponent may be isolated for improvement after its target operational parameters' levels have been determined by the complete supply chain.

Conclusions
Supply Chains and its management pose new challenges through its complexity and dynamics.Existing problem-solving and -evaluation tools may be applied to various dimensions of the supply chain, while simulation-optimisation may become a fundamental tool in supply chain management.
Although dedicated software packages for supply chain management already exist while other similar products are still under development, it was illustrated that

Figure 1
Figure 1 Planning horizons and their characteristics

Figure 3 Figure 3
Figure3shows the product flow and the order flow within the supply chain .Suppliers provide raw material to their downstream manufacturer, while, typically, the information flow is going upstream in the supply chain.With

Figure 5
Figure 5 Flows in the "Customer" Arena module

Figure 9
Figure 9 Effect of the 8M on the lead time Figure 10 Effect of longer processing time

Figure 11
Figure 11 Effects of cyclic demand future is uncertain.whileoptions exist to open or close manufacturing sites, any required capital may be procured and product deployments are open.Ie Large cost decisions are made on this level.rracticalplanning ~Intermediate time horizon (typically several months).~Scope:Regional, on the aggregate level, typically for a family ofproducts.~Thelocation of some facilities are fixed while capacities may be known.IeMany resources are open to adjustment.~There is usually some certainty over the medium term.Limited.,typicallyfor one plant at the part number level.•Resources are known and fixed., with planning aimed at the machining, dispatch, maintenance and material handling.