Simulation models in logistics represent an attempt to replicate mathematically the functional relationships between the logistical activities of facility location, transportation, inventory, order processing and material movement. The usefulness of this replication is realized by experimentation and analysis through manipulation of the simulation model. This allows the manager to study the effects of different managerial policies and environmental conditions upon the logistics system without any manipulation of actual logistical operations.
From a management and research perspective there are many benefits to the use of computer simulation in logistics. For example, the effects of certain decisions can be evaluated before implementation and systems of great complexity can be reduced to a more simple structure for ease in understanding. However, there are also certain problems associated with simulation in logistics:
- the simulation models may be too complex for managerial use;
- there may be little resemblance between the actual system and the model
- the simulation model may be limited in its abilities to analyse the system.
Any simulation model designed for analysis for business situations should meet certain criteria. A possible list of such criteria includes the following
- Minimum of data input.
- Ease of data input.
- Ease of updating.
- Ease of use.
- Ease of understanding.
- Minimum of outside help.
- Minimum of computer cost.
- Rapid turnaround.
- Transferrable software package (hardware compatibility).
- Model a wide range of corporate systems.
- Model a wide range of environmental factors.
- Quantitative analysis.
- Readable, informative reports.
- Educational potential.
The first eight criteria deal with the ease of use of the computer model. The last six criteria deal with its flexibility in application to a variety of business situations. One computer simulation technique which meets most of the above criteria and seems quite suited to logistical problems is the Graphical Evaluation and Review Technique [GERT). Though GERT has experienced many previous managerial applications, it has not previously been utilised in the area of logistics.
Weigel and Cao applied GIS in conjunction with Operations Research [OR) techniques to solve technician dispatching and home delivery problems at Sears, Roebuck and Company. Sears used a vehicle routing and scheduling system based on a geographic information system to run its delivery and home service fleets more efficiently. Although the problems to be solved can be modelled as vehicle routing problems with time windows [VRPTW), the size of the problems and, thus, practical complexity make these problems of both theoretical and practical interest. The authors constructed a series of algorithms, including an algorithm to build the origin and destination matrix, an algorithm to assign resources, and finally algorithms to perform sequencing and route
improvement. The combination of GIS and OR techniques improved the Sears technician dispatching and home delivery business. It
- reduced driving times by 6%,
- increased the number of service orders each technician completed per day by 3%,
- reduced overtime by 15%
- helped to consolidate routing offices from 46 to 22,
- achieved annual savings of $9 million.
The success of this application also suggested a promising link between GIS and OR techniques. It also helped ESRI, the GIS consultant for the project develop ArcLogistics, a low-cost PC-based routing-and-scheduling application that brings high-end functionality to small organizations who were previously unable to afford this technology.
Logistics Simulation in the Process Industry
Although simulation of logistic systems in the process industry has a long lasting tradition it was always difficult to establish the methodology there. The awareness of simulation in logistics in the process industry was not strong enough and obviously the challenges in this industry were elsewhere but the situation is changing and logistics and production aspects are last resorts where money can be saved.
On the other hand, simulation has always had difficulties being applied correctly in the process industries such as: the food industry, chemical industry, pharmaceutical industry, animal care etc. The difficulties are based on the discrete nature of the methodology where liquids or powder is pouring through tanks, silos and tubes and is finally filled in silos, big bags or sacks and transported
and stored in barrels and pallets. An additional complexity is due to the fact that, because of flexibility in operations, control rules tend to be very complicated. Furthermore, many systems’ performances are dependent on the scheduling of processes, equipments, tools, resources, etc.
The potentials, however, are considerable. The range of assets has to be standardized, inventory levels have to be decreased, control rules have to be standardized, product quantities have to be harmonized along the logistic chain and production facilities have to be utilized as much as possible.
The process industry is adding complexity to the simulation studies because it covers nearly all manufacturing aspects and adds process-based needs to the investigations. If it were possible to appropriately fulfill the requirements of the process industry some other application areas such as traffic or mass production (e.g. cigarettes, the textile industry) could follow and benefit from the developments.
