6.3 Differences 
Integrity of agricultural production system should not 
be confused with reliability (the probability of 
performing a function for a period of time), quality( 
customer satisfaction), or availability (ratio of up time 
to total time). Integrity is a higher level measurement 
of performance of agricultural production system that 
indicates the agricultural production system's ability to 
operate in the presence of environmental disaster. 
Architects and desighers can use redundancy to build 
integrity into agricultural production systems. They 
have recognized the critical role that agricultural 
production facilities play in society and the 
consequences of agricultural production system failure. 
6.4 Analysis Approaches 
There are two approaches to survivability analysis 
related to environmental disasters. The method of 
probability models (Barlow and Proschan, 1975; Larson 
and Shubert, 1979) can be used. The first approach uses 
probability of failures of agricultural production system 
and, possibly, rates of repair and restoration to calculate 
various probabilistic measures of availability or 
“unavailability of agricultural production system. The 
second is a conditional approach, defining measures of 
an agricultural production system after given failure 
event have occurred. This approach may either use 
probabilistic weighting of the resulting states of the 
agricultural production system and resulting restoration 
and repair of agricultural production system after the 
failure or use deterministic analysis of these states. The 
approaches can be used to evaluate different 
restoration, repair, or preventive policies. 
7. CONCLUSIONS 
In summary, this paper discusses mainly basic 
principles and key techniques of policy hierarchies for 
environment disaster management. The approach has 
been developed to study the performance of agricultural 
production system. The policy hierarchy concept and its 
application are two keys to design and implement the 
management system of environmental disasters. Thus, 
the approach is also policy-oriented analysis approach. 
The above discussion shows that the usefulness of such 
approach for analyzing survivability of agricultural 
production system and the ability to construct the 
relevant analysis model are limited not by computer 
technology but by knowledge of the model dynamics 
and the effect of various decision upon them. It is that a 
model's dynamics may depend more on the manner in 
778 
which its elements are described and linked than on the 
form of its equation and sub-systems. 
The work to develop more efficient approach to 
management of environmental disaster in agricultural 
production system will become a future challenge in the 
domain. The survivability analysis of agricultural 
production system related to environmental disaster 
will change from the current simple model into future 
more complex one. Many current mathematical tools 
will provide the powerful supports for the challenge or 
objective. However, the future major research works 
about the policy hierarchy are to solve some key 
problems that include how to improve the architecture 
of policy hierarchy model, how to construct the 
operation rules used in the architecture, and how to 
design and implement these rules. In summary, the 
above successful experiences have shown that our 
current work and outcomes provide a satisfactory 
ground and open a wider research domain for the 
development of future survivable agricultural 
production system by policy hierarchy approach. 
8. ACKNOWLEDGMENT 
The authors wish to acknowledge the valuable 
contributions from Mrs. Chen Gui-Zhen to the 
research. This work was supported by both the Project 
of China National Foundation of Natural Sciences 
(Project Number 39470415) and the Project of Navy 
Research. 
9. REFERENCES 
Barlow, R. E. and F. Proschan (1975). Statistical 
Theory of Reliability and Life Testing: Probability 
Models. Holt-Rinehart, Winston, New York. 
Larson, H. J. and B. O. Shubert (1979). Probabilistic 
Models in Engineering Sciences. John Wiely & 
Sons, New York. 
Neumannm, P. (1992). Risks to the public in computers 
and related systems. ACA Software Ing. Notes, 
Vol. 17, no. 1, pp. 3-32. 
Moffet, J. D. and M. S. Sloman (1991). The 
representation of policies as system objects. 
SIGOIS Bulletin, Vol. 12, no. 2 and 3, pp.171- 
184. 
Wu, T. -H (1992). Fiber Network Service 
Survivability. Artech House, New York. 
Intemational Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998 
Du 
col 
firs 
ye: 
me 
dec 
dif 
hy 
sat 
vel 
ab: 
pel 
clo 
rer 
ne 
an 
res 
bu 
sat 
fie 
Or 
kn 
ant 
inf 
ba