Distribution Logistics in Food Industry: a case study
Francisco José Negreiros Gomes
Marcos José Negreiros Gomes
Guilherme Siqueira Simões
Renato Machado Albert
In this paper we describe the situation analysis of the distibution logistics carried out by a medium size food distribution center. First, we present the funcional model of the distribution process in which it becomes clear the need of a new approach to improve the logistics operation. The proposed methodology consists of a geo-referenced distribution and transportation system that interacts witha storage management system to be aware of the food stock levels and with the integrated management system, applying forecasting tools.
Logistics, GIS, Transportation, Integrated Systems
In the supply chain management, the logistics eviroment is contantly ecolving as a result of changes in markets, competitor, suppliers, and technology. Inorder to develop or refine an enterprise strategy to match this changing evironment, a systematic planning and design methodology is necessary to formaly include the relevant consideration and effectively evaluate alternatives. In this work, we provide a general approach for distribution logistic planning an present a case study in the food distribution.
2. FUNTIONAL MODELLING
Tool for Business Process Modelling
Modern logistic system studies consider business process identification and analysis the central point of their approaches. To study business process the very simply and well-known techniques, such as flow charts and data flow diagrams are the tools of common preference. However these techniques are seriously inadequate when used for the purpose they have never been developed for, Zgorzelski and Zagorzelsa[ ].
As an alternative to the "standard"flow charts and data flow diagrams, the IDEF has been adopted as a language of manufacturing system description. The origins of the IDEF family of system analysis and design techniques lie in the ambitious U.S. Air force program called Integrated Computer Aided Manufacturing undertaken in the 1970's.
IDEF is a standard, public-doamin modeling language that has been succefully used for over twenty years. IDEF diagrams use simple rules to govern graphical layouts of a system or process, giving the technique as intuitive appeal and making it easy to learn. IDEF uses functional decomposion to provide both a big picture and a detailed view of the process(es) being analysed. The method's notation encourages simple models that generalizes as much as possible at the higher levels while providels as many levels of low level details as necessary. For these reasons, IDEF serves not only as a useful modeling tool but also as an excelent communication tool.
IDEF uses a simple visual vocabulary that has been carefully designed for describing complexes process in sufficient detail. Boxes are used to represent activities. Arrows from left an to the right show the flow of information into and out of an activity. Arrows from beneath a box show the resource required by activities. The IDEF notation also uses the triggers that are involved in the control of a process. The two aspects, triggers and resources, make IDEF a powerful tool for decribing when and how things happen, an essentiio\al element for complex analysis such as business process design.
Modelling the Logistical Aspects of the Distribution Company
The main task of the distribution company that we study is the
trade of fruits and other food products. Figure 1 shows the
primary IDEF diagram of the business process. The diagram
constitutes of the three main processes of the company regardiing
to distribution logistics. It starts with the purchasing and
transportation of fruit and other food products from the
supplyers through buyers. The quantities to be purchased are
estimated from the information provided by the stock managers and
the back orders. The second process is concerned the storage of
the goods which takes place in three sites (company's stores, the
transportation department and the frozen warehouses). These sites
mantain the flow of goods to support the commercialization and
the distribution of the food products. The final process is
related to the slling and distribution of food products. This
process is activated by the costumers' orders and it is carried
out by the company fleet of trucks to distribute the goods which
may be picked up from the warehouses or direct from the stores.
The previous diagram can be detailed into a low level view as showed by Figure 2. The next diagram describes the process of buying and tranportation of fruits and other food products. This process can be divided into three sub-process. First the orders took place from the buyers because of the stores position and estimated demands. Second, the orders are transmitted to the suppliers . Note that goods are either delivered directed to the stores or are delivered by the transportation department. The third part of this process is the actual transportation of goods using the company's fleet.
An Instantaneous Analysis
As we can see by the diagram of figure 2, the flow of goods goes from the warehouses to the stores, and then fom the stores to the costumers. This part of the process increases very much the handling and transportation cost. The diagram helped us to identify a basic weakness of the business process: the unecessary transshipment of the goods. Certainly, it would be better to send the orders to the transportation department, and then it should distribute the goods direct to the customers.
3. TECHNICAL MODELLING OF LOGISTICS
The objective of modelling the distribution company supply chain with respect to the logistics aspects is to find the right balance between cutting costs and mantaining costumer responsiviness.
In this way, the tasks that will be carried out are the following: identify, analyse and recommed the most cost-effective changes for warehousing and distribution of their goods.
The approach we propose is to use a Supply Chain Simulation/Optimization model - SCSOM that will help to compare different planning and forecasting methods and to identify lower cost distribution policies. In the analysis phase of the modelling, the first step should be to understand the warehousing and distribution functions. Second, it would necessary to identify the key ares of cutting costs and custumer responsiviness.
The modelling, as per the proposed methology of the project, will be based on object oriented language, with support for simulation, together a business process modelling tool. The modelling of system objects will be represented as nodes and arcs in order to capture the behavior of various business functions and to model the movement of goods and information.
3.1 Logistical Redesign Process
A general process applicable to most loigistics design and analysis situations could be structured as follows:
Phase I: Problem Definition and Planning
Phase II: Data Collection and Analysis
Assumptions and Data Collection
Phase III: Recommendation and Implementation
Development of Recommendations
3.2 Definition of DSS Requirements
The desired funtions and features for a Decision Support System in the context of logistics should include the analysis and evaluate capabilities. A desired feature list defines the data entry, manipulation, and reporting capabilities of the system. Below, we provide an extensive, although not necessarily exhautive, list of features that should be cosidered for logistics DSS. Input from funcional managers, information systems personnel, and experienced users should be otained when defineing requirements.
1. Data Entry. Can accept data from common spreadsheet and GIS packages.
2. Data Manipulation. Facilitates data manipulation to complete "what if"analysis.
3. Scope. Consider relevant logistics costs including inbound transportation, distribution center operations, and customer freight.
4. Search Algorithms. Includes algorithm or heuristics to search for and identify the optimum or best alternative.
5. Reporting. Includes praphical reporting and display capacity
1. Operating requirements. Consideration of the campability of the package with existing hardware, planned hardware, and system software.
2. Design quality. Consider the overall software design and the resulting ease of software modification.
Object Oriented and agent based design techniques make the instalation more esier and flexible.
3. Documentation. Includes current, complete, and cross-referenced technical documentation to expedite modification and installation effort.
Decision Support Requirements
The software should be able to help the user in the decision questions such as follows:
- How should delivers be grouped into routes?
- What is the best deliver sequence?
- Which routes should be assiened to vehicles?
- How will delivery time restriction be imposed?
Answers to the above questions are fundamental to attain the most important distribution logistics costs: Minimize the combination of vehicles, hours, distance requires to deliver food products.
4. A FLEXIBLE DSS FOR TRANSPORTATION LOGISTICS
The design and implementation od a DSS for transportation logistics is a challenge. There are several softwares packages to support fleet managers in the scheduling and routing of vehicles, Hall and Partyka [ ].The specificity of the problems that arises in companies such the one that we analysed make the them difficult to solve using the existing software packages. These packages cannot take into account important business constraints. The necessary customization of these packages, if it is possible, makes the approach very costly. Besides, including such constraints may lead to an intractable problem in the computational sense.
The SisGRAFO system, Negreiros et al (1997), was designed with
the ojective to be an graphical computer environment in the
solution of graph optimization problems.The system incorporates
de Visual Interactive Modelling paradigm and an Algorithm
Catalog.. Figure 3 presents a solution of a in city roting