Interpolation result —_ 
ar 
nterpolation result » 
1.10001 1111 T ses 
Time fitness=1.138443e-02 
Na 
11111000 Space 
Y fitness-1.138443e-02 
Time 
    
    
  
  
  
   
  
  
  
  
  
    
(11100011 
Tim fitness=1.939160e-04 
  
  
  
  
Figure.9 Interpolation Results in Non-overlapping Case 
  
  
  
1111100011 
1111 500011 
  
  
  
  
  
Time 
  
Tim 
Observational data im Interpolation result 
Figure.10 Interpolation Result with 
Additional Observational Data 
6. CONCLUSION AND FUTURE PROSPECTS 
In this study, a spatio-temporal interpolation scheme 
is proposed for raster nominal data which can integrate 
observational data with behavioral/structural models/rules 
under the framework of maximizing likelihood of spatio- 
temporal events. Genetic-Algorithm/Hill-Climbing can be 
successfully applied to the combinatorial optimization of 
nominal voxel-field data. Conclusions from the experiments 
can be summarized as follows: 
1) GA/HC can be very rigorous because it can generate the 
most likely spatio-temporal distribution of class variables 
under observational data and a behavioral model; 
2) Hill-Climbing method can be effective method to greatly 
improve the efficiency of GA; 
Although the GA/HC authors proposed can be a good 
scheme for spatio-temporal interpolation, it is just a first 
attempt to apply GA in the field of spatio-temporal 
interpolation. We will apply GA/HC for larger size of class 
variable data, to reduce the speed of premature convergence 
and get higher efficiency of GA/HC. 
[1] Bramlette, M.F. (1991): Initialization, Mutation and 
Selection Methods in Genetic Algorithms for Function 
Optimization, Proc. of the 4th. Conf. on GA, R.K Belew 
and L.B.Booker (Editors), July, 1991, pp.100-107. 
[2] Booker, L. (1987): Improving Search in Genetic 
Algorithms, Genetic Algorithms and Simulated Annealing, 
edited by Lawrence Davis, Morgan Kaufmann Publishers, 
Inc. Los Altos, California, 1987, pp.61-73. 
[3] Burrough, P.A. (1986): Methods of Spatial 
Interpolation, Principle of GIS for Land Resource 
Assessment, Monographic on Soil and Research Survey, 
No.12, 1989, pp.147-166. 
[4] Davis, L. (1987) : Genetic Algorithms and Simulated 
Annealing, Pitman Publishing, 128 Long Acre, London 
WC2E 9AN, 1987. 
[5] Eshelman, L.J. and J.D.Schaffer (1991): Preventing 
Premature Convergence in Genetic Algorithms by 
Preventing Incest, Proc. of the 4th. Conf. on GA, 
R.K.Belew and L.B.Booker (Editors), July, 1991, pp.115- 
122. 
[6] Gold, C.M. (1989): Surface interpolation, spatial 
adjacency and GIS, Three Dimensional Applications in 
Geographic Information System, Edited by J.Raper, Taylor 
& Francis Ltd., 1989, pp.22-35. 
[7] Goldberg, D.E. (1989): Genetic Algorithms in Search, 
Optimization & and Machine Learning, Addison-Wesley 
Publishing Company, Inc., 1989. 
[8] Olover, M.A. and R.Webster(1990): Kriging: a method 
of interpolation for geographical information systems 
Int.J.Geographical Information Systems, 1990, Vol.4, No.3, 
pp.313-332. 
[9] Shibasaki,R., T.Ito and Y.Honda (1993): Integration 
of Remote Sensing and Ground Observation Data for 
Developing Global GIS, Proc. of SEIKEN Symposium, 
Vol.12, Aug. 1993, pp.263-277. 
[10] Shibasaki,R., M.Takaki and Y.Honda (1994):Spatio- 
temporal Interpolation Using Multi-Source Data for 
Global Dataset Development, COMMISSION III, IGWG 
HI/IV; 1994. 
[11] Huang, S.B. and R. Shibasaki (1995): Development 
of Generic Algorithm / Hill-climbing Method for Spatio- 
temporal Interpolation, Sixth symposium on functional 
graphics and geographic information systems ,IIS, Univ. of 
Tokyo, 1995. 
808 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B3. Vienna 1996 
      
     
    
     
    
  
   
     
    
      
   
    
  
   
   
   
   
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
   
   
   
  
Th 
im; 
del 
ab: 
tor 
COI