影響，讀取器在物流出入口布置又需要做哪些調整；以上這些議題希望能夠在我 們未來延伸的研究中持續進行探討。
參考文獻 參考文獻 參考文獻 參考文獻
[1] Asif, Z., and Mandviwalla, M., "Integrating the supply chain with RFID: A technical and business analysis." Communications of the Association for Information Systems, pp.393-427, 2005.
[2] Balanis,C. A., Antenna Theory Analysis and Design, second edition, John Wiley
& Sons Inc.,1997.
[3] Bret, K., "THE Wal-Mart FACTOR," Industrial Engineer, vol.35 no. 11, pp.32-36,2003.
[4] Calegari,P., Guidec,F., Kuonen,P., and Wagner,D., "Genetic Approach to Radio Network Optimization for Mobile Systems," IEEE VTC, pp. 755-759, 1997.
[5] Croft, H. T., Falconer,K. J. and Guy, R. K.. Unsolved Problems in Geometry.
Springer-Verlag , 1991.
[6] Dobkin , D.M., The RF in RFID: Passive UHF RFID in Practice,Newnes ,2007.
[7] Finkenzeller, K., RFID Handbook: Fundamentals and Applications in Contactless Smart Cards and Identification, second edition, John Wiley & Sons Inc., 2003.
[8] Fisher,N. I., Lewis, T., and Embleton, B. J. J. Statistical Analysis of Spherical Data. Cambridge University Pres, 1987.
[9] Fonseca,C.M.,Fleming,P.J.,"Multiobjective Genetic Algorithms Made Easy:Selection,Sharing and MatingResteiction, Genetic Algorithms in Engineering Systems: Innovations and Applications, " GALESIA,First International Conference, 1995
[10] Gaukler,G.,RFID in Supply Chains,PhD thesis, Stanford University, Stanford,2005.
[11] Greene, C.E., Area of Operation for a Radio-Frequency Identification (RFID) Tag in the Far-Field , Ph.D.Dissertation, Department of Electrical Engineering,
University of Pittsburgh, 2006.
[12] Greene ,C. E., and Mickle ,M., Determining the three-dimensional read accuracy of an RFID tag using a power scaling factor. Working Paper, Dept. of
Electrical and Computer Engineering, 2006.
[13] Han, J. K., Park,B. S., Choi ,Y. S., and Park,H. K., ”Genetic Approach with a New Representation for Base Station Placement in Mobile Communications,”
Proc. IEEE 54th Vehicular Technology Conf., Vol. 4, pp.2703-2707, 2001.
[14] Holland, J.H., Adaptation in natural and artificial systems., MIT Press Cambridge, MA, USA, 1992.
[15] Huang,C.-F., Tseng,Y.-C., and Lo,L.-C., "The coverage problem in three-dimensional wireless sensor networks. ," Proc. IEEE GLOBLECOM '04, volume 5, pp. 3182-3186, 2004.
[16] Lee, Y. M., Cheng, F., and Leung, Y. T., " Exploring the impact of RFID on supply chain dynamics. " Proc. of the 2004 Winter Simulation Conference, vol.
2,pp. 1145-1152, 2004.
[17] Man, K.F., Tang, K.S.and Kwong,"Genetic algorithms: concepts and applications [in engineering design] ," S.Industrial Electronics, IEEE Transactions, vol.43 , no. 5, pp.519-534, Oct 1996.
[18] Meunier,H., Talbi ,E., and Reininger,P., "A multiobjective genetic algorithm for radio network optimization, " Proc. of the 2000 Congress on Evolutionary Computation, Vol. 1 ,pp. 317–324,2000.
[19] Michalewicz, Z., Genetic Algorithms + Data Structures = Evolution Programs, 2nd ed., New York: Springer-Verlag, 1994.
[20] Michael, K. and McCathie, L., "The pros and cons of RFID in supply chain manage-ment. " Proc. ICMB'05, pp. 623-629, 2005.
[21] Mitchell, M., An Introduction to Genetic Algorithms. , Cambridge, MA: MIT Press, 1996.
[22] Niemeyer, A., Minsok, H.P. and Ramaswamy, S.E., "Smart tags for your supply chain.,"McKinsey Quarterly,pp.6-9,2003.
[23] Rajkumar,A., Naylor,B., Feisullin,F., and Roger,L., "Predicting RF coverage in largeenvironments using ray-beam tracing and partitioning tree represented geometry. ,"Wireless Networks, pp. 143-154, 1996.
[24] Rao,K.S., "An overview of backscattered radio frequency identification system(RFID) ". IEEE APMC,vol 3, pp. 746-749,1999.
[25] Reininger,P., Iksal, S., Caminada, A., and Korczak ,J., "Multi-stage optimization for mobile radio network planning. ," Proceedings of the IEEE VTC 99, pp.
2034-2038, 1999.
[26] Rusin ,D., Topics on sphere distributions,
URL:http://www.math.niu.edu/~rusin/known-math/95/sphere.faq.(Accessed at:2009.11.19)
[27] Sa , E. B. and Kuijlaars, A. "Distributing many points on a sphere. "
Mathematical Intelligence, 5-11, 1997.
[28] Tabbane,S., "Location management methods for third generation mobile systems,
" IEEE Communication Magazine 35, vol. 8, pp. 72-78, 83-4, 1997.
[29] Wang, L., Norman,B. A. and Rajgopal,J., "Placement of Multiple RFID Reader Antennas to Maximise Portal Read-Accuracy , " International Journal of Radio Frequency Identification Technology and Applications, vol. 1, pp. 260 - 277, 2007.
附錄 附錄
附錄 附錄 基因演算法 基因演算法 基因演算法 基因演算法
輪盤法 輪盤法 輪盤法 輪盤法
for e =1:e_time
clear all;
popusize = 20 generation = 30 xover_rate = 0.3 mutation_rate = 0.05
init_popu=gen_init_popu(popusize);
popu_fit = fit_check(init_popu,popusize);
best_popu = bestpopu_check(init_popu,popusize,popu_fit);
new_popu = gaoperator( popusize , init_popu , popu_fit , xover_rate ,best_popu );
for i = 1:generation
rand('state',sum(100*clock));
new_popu_fit = new_fit_check(new_popu,popusize);
best_newpopu = bestnewpopu_check(new_popu,popusize,new_popu_fit);
new_popu = new_gaoperator( popusize , new_popu , new_popu_fit , xover_rate ,best_newpopu );
best(:,i) = max(new_popu_fit(:,2));
middle(:,i) =sum(new_popu_fit(:,2))/popusize;
%median(new_popu_fit(:,2)) ;
worest(:,i) = min(new_popu_fit(:,2));;
if rand < mutation_rate
new_popu = mutation(popusize, new_popu);
end
end
figure
x = (1:generation)';
plot(x, best, 'o', x, middle, 'x', x, worest, '*');
hold on;
plot(x,best,x, middle,x, worest);
hold off;
legend('Best', 'Average', 'Poorest');
xlabel('Generations'); ylabel('Fitness');
plot(x, best,':go' ,x, middle,'-b*', x, worest,'-ro');
legend('Best', 'Average', 'Poorest');
xlabel('Generations');
ylabel('Fitness');
end toc
競爭法 競爭法 競爭法 競爭法
yes = 1;
no = 0;
global print_message_screen;
global print_message_file;
print_message_screen = yes;
print_message_file = no;
print_result_file = yes;
if print_message_file == yes
global fid;
fid = fopen('portal_log20090510.txt','w');
end
global generation_n;
generation_n =30; % Number of generations popuSize = 20; % Population size
xover_rate = 0.3; % Crossover rate mutate_rate = 0.05; % Mutation rate
bit_n = 11; % Bit number for each input variable var_n = 1; % Number of input variables
experiment_times = 25;
global reader
reader =[0, 0.5, 1.5, 90, -45; 0, 0.5, 1.5, 90, 0; 0, 0.5, 1.5, 90, 45;
0, 1, 1.5, 90, -45; 0, 1, 1.5, 90, 0; 0, 1, 1.5, 90, 45;
0, 1.5, 1.5, 90, -45; 0, 1.5, 1.5, 90, 0; 0, 1.5, 1.5, 90, 45;
0, 2, 1.5, 90, -45; 0, 2, 1.5, 90, 0; 0, 2, 1.5, 90, 45;
0, 2.5, 1.5, 90, -135; 0, 2.5, 1.5, -90, 0; 0, 2.5, 1.5, 90, -45;
0.5, 3, 1.5, 90, -135; 0.5, 3, 1.5, 90, -90; 0.5, 3, 1.5, 90, -45;
1, 3, 1.5, 90, -135; 1, 3, 1.5, 90, -90; 1, 3, 1.5, 90, -45;
1.5, 3, 1.5, 90, -135; 1.5, 3, 1.5, 90, -90; 1.5, 3, 1.5, 90, -45;
2, 3, 1.5, 90, -135; 2, 3, 1.5, 90, -90; 2, 3, 1.5, 90, -45;
2.5, 3, 1.5, 90, -135; 2.5, 3, 1.5, 90, -90; 2.5, 3, 1.5, 90, -45;
3, 2.5, 1.5, 270, 45; 3, 2.5, 1.5, 270, 0; 3, 2.5, 1.5, 270, -45;
3, 2, 1.5, 270, 45; 3, 2, 1.5, 270, 0; 3, 2, 1.5, 270, -45;
3, 1.5, 1.5, 270, 45; 3, 1.5, 1.5, 270, 0; 3, 1.5, 1.5, 270, -45;
3, 1, 1.5, 270, 45; 3, 1, 1.5, 270, 0; 3, 1, 1.5, 270, -45;
3, 0.5, 1.5, 270, 45; 3, 0.5, 1.5, 270, 0; 3, 0.5, 1.5, 270, -45];
exper_result = zeros(experiment_times, 3); % 1st col max value, 2nd col max vale happen generation
for k = 1 : 1 :experiment_times
% Initial random population
popu = rand(popuSize, bit_n*var_n) > 0.5
upper = zeros(generation_n, 1) average = zeros(generation_n, 1) lower = zeros(generation_n, 1)
max_fit_gen=0;
max_fit_value=0;
% Main loop of GA
for i = 1:generation_n;
if print_message_screen == yes
fprintf('---Generation %d---\n', i);
end
if print_message_file == yes
fprintf(fid, '---Generation %d---\n', i);
end
% delete unnecessary objects
delete(findobj(0, 'tag', 'member'));
delete(findobj(0, 'tag', 'individual'));
delete(findobj(0, 'tag', 'count'));
% Evaluate objective function for each individual fcn_value = evalpopu(popu, bit_n);
for j = 1:popuSize
if print_message_screen == yes
fprintf('Population %d is %f and %f\n', j, fcn_value(j,1),
fcn_value(j, 2));
end
if print_message_file == yes
fprintf(fid, 'Population %d is %f and %f\n', j, fcn_value(j,1), fcn_value(j, 2));
end end
% Fill objective function matrices upper(i) = max(fcn_value(:,2));
average(i) = mean(fcn_value(:,2));
lower(i) = min(fcn_value(:,2));
if print_message_screen == yes
fprintf('upper = %f, average = %f, lower = %f\n', upper(i), average(i), lower(i));
end
if print_message_file == yes
fprintf(fid, 'upper = %f, average = %f, lower = %f\n', upper(i), average(i), lower(i));
end
% display current best
[best, index] = max(fcn_value(:, 2));
exper_result(k, 1) = best;
if best > max_fit_value max_fit_value = best;
exper_result(k,2) = i;
end
exper_result(k, 3)=fcn_value(index, 3);
% generate next population via selection, crossover and mutation popu = nextpopu(popu, fcn_value(:, 2), xover_rate, mutate_rate);
end % Main loop of GA
[best, index] = max(fcn_value(:, 2));