OPTIMAL MANAGEMENT SYSTEM OF STORED GRAIN IN GIS ENVIRONMENT

This paper describes the development of an optimal management system of stored grain in GIS environment ,which provides a new method for designing and implementing an advanced stored grain management system for farmers and specialists related to stored grain management. In the project, the general principle and basic method of designing the system were developed. The system consisted of key subsystems such as expert system, simulation models and basic GIS environment. Especially, Esri�s advanced GIS software ARCVIEW 2.1 was used to integrate all subsystems into the final system. A number of simulation models were developed using the numeric or graphic variable methods under ARCVIEW environment, which are used to monitor and describe some dangerous status such as dangerous temperature and moisture values in stored grain. A layered architecture model was developed to describe architecture of the stored management system and to guide the system design and implementation. An ARCVIEW real world based on the architecture was built to complete final O-O implementation of the system. Some key techniques such as design of various basic models, their each other relationship analysis and how to integrate them under ARCVIEW environment were also discussed in detail. The performance of prototype of this system is very satisfactory. In current time, we have been testing and refining the prototype functions in order to constant larger practical application system based on it.

1. INTRODUCTION

Scientific management of stored grain is one of the most parts that 

consist of grain production system in modern agricultural production. 

In recent several years, the number of stored grain loss during the 

store period in the world has been increasing due to the very poor or 

no scientific management of stored grain, especially in most of 

developing countries. Thus, it is really necessary to construct an 

advanced management system of stored grains with some current 

advanced techniques of information science for implementing their 

scientific management in order to increase modern level of grain 

production. In fact, because of the possible large size of grain shortage 

in the world extent in next century, the constructions of advanced 

stored grain facilities and related management systems will become a 

more and more problem in the future. However, how to design and 

construct the management system has been being a very complex and 

important problem in the researches on stored grain management 

engineering. In past many years, a number of technical papers had 

discussed deeply the problems how to design and implement a number 

of stored grain management systems and provided many usual 

practical experiences and outcomes. But, we consider that those 

outcomes lack a scientific system method to describe and solve this 

problem, so that so far is how to implement the scientific management 

of stored grain still a very complex and difficult problem. In general, 

this stored grain management belongs to a problem of large complex 

system management about optimal control and decision that involves 

other problems about how to control and manage many complex 

objectives. Thus, there are a number of  factors to influence the design 

and implementation of this system, which include the stored grain 

kind and quality, stored grain facilities, environmental temperature 

and moisture, and various possible pests, etc..  Main objective of this 

management is to control the stored grain environment using natural 

and artificial aeration means to destroy any possible conditions 

resulting in any possible stored grain disasters in order to assure that 

temperature and moisture of stored grain are in the safety range.

In general, an advanced stored grain management system based on a 

green control rule (no chemical means) must complete two basic tasks 

as follows.

(1)Estimate or predict the temperature and moisture values of any 

position in a stored grain entity in any necessary time.

(2)Manage or control various possible dangerous states of the 

temperature and moisture in the entity  for storing grain because they 

could result in various possible stored grain disasters.

One of most important tasks completed by the advanced grain 

management system is to precisely predict the temperature and 

moisture in the grain of all stored grain entities of grain warehouses 

distributed on a very large geographical areas. The previous or past 

mathematical models developed  for completing the task used mainly a 

linear regression concept to predict the temperature and moisture and 

did not consider the strong influence from geographic region factor or 

information to the prediction. Further, because those models used only 

a few factors effecting on the temperature and moisture as their input 

parameters, and considered that all physical positions in a stored grain 

entity had the same temperature and moisture value, so that their 

precision and practical value in practical application were not 

satisfactory. However, in recent several years, some mathematical 

models based on the analysis of differential equation and vector field 

had be developed to predict the temperature and moisture on any space 

position in stored grain entity on the condition that the stored grain 

entity was investigated as a complex temperature and moisture fields. 

Some of those models could be used to predict the temperature and 

moisture on any spatial point in the entity, whose precision was 

basically satisfactory. Otherwise, those models have also some 

common and obvious disadvantages such as large computation cost, 

vague computation results and obvious difficulty to determine various 

dangerous states. Further, the main reason resulting in the above 

disadvantages is really that those models belong usually to one 

dimension numeric computations model. However, our research shown 

that those models must be reasonably improved to overcome their 

disadvantages. On the other hand, a modeling method based on 

graphic variables had be successfully developed and be used to design 

and implement the mathematical model used in our advanced stored 

grain management system. This paper presented mainly our new 

method for designing and implementing an optimal management 

system of stored grain in GIS environment and our new advances in 

the research of stored grain management. First, the general principle 

and basic method of designing basic model of the management system 

were discussed. Secondly, the basic GIS environment based on 

ARCVIEW 2.1 and how to implement the above basic model in it 

were respectively investigated. Thirdly, some key techniques which 

were used to implement the management model in the GIS 

environment were discussed, which includes how to integrate the 

optimal system from several subsystems.



2. MANAGEMENT SYSTEM ARCHITECTURE.

The architecture of optimal stored grain management system can be 

described by a seven-layer    architecture  model  shown in  Fig. 1. The  

layer  of  basic   facilities 

describes the function, geographic position, performance and quality 

of basic facilities for the management of stored grain. The layer of 

stored grain kinds describes all information for management of kinds 

of stored grain, which includes the number of kinds of stored grain 

and the characters of stored grain in the facilities. The third layer is 

weather environment layer, which is used to implement weather 

environment management of stored grain. In the management, a 

number of historical and current weather data are used to analyze the 

weather environment of regions where stored grain facilities are built 

and managed. The main task of decision data layer is to generate a 

number of basic data form service of the low three layers for 

supporting operation of decision models in its high layer. The decision 

layer consists of a number of basic decision models to make basic 

optimal management decisions of stored grain. The management layer 

is used to define various different integrated management decisions for 

different stored grain facilities according to practical requirement of 

those facilities and with the help of various basic decision services 

provided by decision layer. The application layer is mainly to support 

various different applications of optimal stored grain management in 

order to suite the needs from various different users. In the layered 

architecture, the lower layer provides always its service for its higher 

layer, the higher layer uses the service from its lower layer. Our 

experiences have shown that to construct the layered architecture 

model is a very effective method for designing and implementing a 

high quality GIS application system.



3. ARCHITECTURE IMPLEMENTATION

The above layered model can be further divided into three different 

subsystems such as basic management, advanced management and 

application subsystems. The basic management subsystem provides 

mainly some raw management data about stored grain facilities, kinds 

and weather environment and some simple management suggestion 

for the management system users. The subsystem structure and 

relationship with GIS environment is shown in Fig. 2. Obviously, 

operation of basic management subsystem is supported mainly by GIS 

environment. This is because three management tasks of the subsystem 

have very close relationship with geographic information. For 

example, a number of data used to describe weather environment, 

managed grain kinds and various basic facilities about stored grain 

have very obvious geographic characters. Thus, this subsystem is 

finally implemented in GIS environment. Otherwise, this subsystem 

provides also various services for advanced management subsystem 

with the exchange of basic management data between them under 

custom/server (c/s) mode.

Three layers on basic management subsystem construct the advanced 

management subsystem. This management subsystem is mainly used 

to complete all optimal control and management works about stored 

grain according to basic management data from its next subsystem. A 

typical advanced management subsystem is shown in Fig. 3. The 

decision data layer makes mainly use of various simulations models 

further generate a number of support data that are used by decision 

models to provide optimal control and management decision for stored 

grain. These models can further be divided into two classes such as 

numeric simulation models and graphic simulation models in order to 

suite different requirements from various different decision models. In 

fact, the numeric models are very common and support mainly 

operation of numeric decision models. However, graphic simulation 

models are used to provide graphic decision data for our new graphic 

decision models in the decision layer. The operation of graphic 

simulation model was implemented on GIS processor based on GIS 

environment with the help of complex operations among various 

graphic coverages about the data from basic management subsystem. 

In the decision layer, an expert system is used to support operation of 

those decision models, which is a rule-based system containing an 

inference engine, a file maintenance system for the decision model 

input data, a database system for the knowledge base, and a interface-

driven system for interactions of higher and lower layers. The 

inference engine applies rules to set up weather and management 

practices, to execute a special decision program, and to interpret the 

operation results of decision models to make recommendations on the 

control and management of temperature, moisture and pest of stored 

grain. In the expert system, the rule-base is based on some excellent 

expert experiences and ability to run decision models, interprets their 

results for higher layer and make the layer produce various effective 

management decision. There are a number of application-oriented 

management models in the management layer, which are designed and 

implemented in order to flexibly satisfy the requirement of various 

typical stored grain facilities.

The application subsystem includes only the application layer, which 

is shown in Fig. 4. There are three classes of models to support the 

subsystem operation. The 

local application models describe various different requirements and 

application means of local users when they use the optimal 

management system of stored grain. On the other band, the remote 

application models provide a number of means to support those remote 

users to use the management system though computer network or any 

other telecommunication network. The models for testing system 

functions are designed and implemented for determining whether the 

management system can complete its predetermined various 

operations. Otherwise, those models assure that the management 

system can complete its normal operation. In fact, our practices have 

shown that the layered architecture model and the division of those 

subsystem as well as the implementation of those subsystems are very 

reasonable for designing and implementing the optimal management 

system of stored grain in GIS environment.



4. IMPLEMENTATION BASED ON ARCVIEW 2.1

According to the above discussion, the optimal management of stored 

grain is a very complex system related closely to geographic 

information and includes several relative alone subsystems. Thus, how 

to integrate those subsystems become really another key problem to 

determine whether the management system can obtain its success in 

practical applications. We selected Esri�s advanced GIS software 

ARCVIEW 2.1 with Avenue as a basic frame work to implement the 

above integration. The relationship of three subsystems and 

ARCVIEW is shown in Fig. 5. This figure shows that the three 

subsystems make use of ARCVIEW and a 

number of scripts developed by Avenue finally implement their 

integration. During the practical integration for those subsystems, 

those subsystems and ARCVIEW construct logically a simple star 

network where ARCVIEW is used as a switching node to implement a 

number of data exchanges among those subsystems, which is shown in 

Fig. 6. The star structure describes in practice a object-oriented design 

and integration decision of the optimal management system of stored 

grain. From the star network, I can construct an ARCVIEW world to 

integrate all key modulars in those subsystems. This connection 

among those subsystems is facilitated by ARCVIEW or scripts 

developed by Avenue that translates data and procedure calls between 

subsystems and also controls execution of the entire system. The O-O 

design based on the star network does not allow subsystems to 

communicate directly. All communication is governed by the 

ARCVIEW real world model of stored grain application domain. The 

advantage of this method was that the management system would not 

be built around any specific type of subsystem, so as to make the 

system compatible with many more environments and applications. 

ARCVIEW real world model of Fig. 7 shows how to integrate all 

subsystems and 

implement the O-O design. The ARCVIEW with Avenue, VISUAL 

BASIC 4.00 and VISUAL C++ 2.0 were used to develop various 

scripts, objects, programs and procedures for implementing the 

integration. In practical implementation of the system, we found that 

ARCVIEW 2.1 and VISUAL BASIC 4.0 provided very satisfactory 

user interfaces for our system, on the other hand, ARCVIEW 2.1 

completed also some complex computation about graphic coverges for 

supporting the operation of graphic simulation and decision models. In 

the process to develop and implement the system, VISUAL C++ 2.0 

was used to design various complex objects and to complete various 

complex numeric computation and simulation operation. Otherwise, 

our current research and achievements had shown that ARCVIEW 2.1 

with Avenue is one of most efficient and flexible environments or 

frameworks for developing the optimal management system of stored 

grain.



5. DISCUSSION

The prototype of optimal management system of stored grain was 

validated using a number of experimental data collected from various 

sources. In addition, the prototype system has been extensively 

validated by our team, who compared the system�s predictions and 

decisions with some field data collected at several sources. The final 

results for all resources was very satisfactory and with the 

predetermined performance range. The system is currently being 

evaluated technically. It is necessary to determine economic value of 

the system in application of practical stored grain management and to 

further develop simulation models to aid the analysis of complex 

stored grain environments. Further research is necessary to look into 

means of dangerous state prediction of stored grain in GIS 

environment, which is the most difficult area of the research project 

because of the high computation cost and respondent time of the 

current system. Further study is required to look into appropriate ways 

of increasing the precision of graphic simulation and decision models 

and decreasing their computation cost where Esri ArcInfo�s 3D 

function will be used to support the development.



6. ACKNOWLEDGMENTS

The authors wish to acknowledge the valuable contributions of other 

members of our research team to the research. They also acknowledge 

funding of the research by China National Foundation of Natural 

Sciences.



7. REFERENCES

Harrison, S. R., 1991. Validation of agricultural expert systems. Agric. 

Syst., 35: 265-286.

Rumbaugh, J., Blaha, M., Premerlani, W., et. Al., (1991) Object-

Oriented Modeling and Design. Prentice-Hall, Englewood Cliffs, NJ, 

500 pp.

Turban, E. (1992) Expert systems and applied artificial intelligence. 

Macmillan, New York.

Power, J. M., 1993. Object-Oriented design of decision support 

systems in natural resource management. Compute. Electron. Agric., 

8: 301-324.

Lob, D. K., et al ., 1994. Integration of a rule-based expert system with 

GIS through a relational database management system for forest 

resource management. _, 11: 215-228.

Chang, C. S. And J. L. Steele. 1995. Development and evaluation of 

aeration control straggles for maintaining stored grain quality. 

Applied Engineering in Agriculture 11(4): 577-582.

Chang, C. S., H. H. Converse and J. L. Steele. 1994. Modeling of 

moisture content of grain during storage with aeration. Transactions of 

the ASAE 37(6): 1891-1898.

Chang, C. S., H. H. Converse and J. L. Steele. 1993. Modeling of 

temperature of grain during storage with aeration. Transactions of the 

ASAE 36(2): 509-519.

Singh, A. K., E. Leonardi and G. R. Thorpe. 1993. A solution 

procedure for the equations that govern three-dimensional free 

convection in bulk stored grains. Transactions of the ASAE 36(4): 

1159-1173.