A STOCHASTIC-MODEL OF LUMPING

被引:1
作者
IORDACHE, O [1 ]
CORBU, S [1 ]
机构
[1] POLYTECH INST,DEPT MATH,R-78126 BUCHAREST,ROMANIA
关键词
MATHEMATICAL MODELS - MIXERS;
D O I
10.1016/0009-2509(87)80215-4
中图分类号
TQ [化学工业];
学科分类号
0817 ;
摘要
A natural procedure to simplify the study of a large system is the lumping of states in classes. Here the states of a Markov chain associated with a compartmental model of mixing are grouped in classes. Our aim is to study the non-exactly lumpable systems in which the partition in classes is imposed by a economic and experimental constraints. In such a case the lumped chain is non-Markovian. Consequently RTD must be replaced by conditional RTD. Some ways of obtaining conditional RTD by simple experiments in which only the classes itinerary is recorded are presented. Different lumpings are compared using the entropy as a measure of loss of information by lumping.
引用
收藏
页码:125 / 132
页数:8
相关论文
共 15 条
[1]  
Aoki M., 1967, OPTIMIZATION STOCHAS
[2]  
Blackwell David, 1957, 1 PRAG C INF THEOR S, P13
[3]  
CUCULESCU I, 1984, PROBABILITIES STOCHA, P100
[4]  
FAN LT, 1982, RESIDENCE TIME DISTR, P75
[5]   SOLUTION OF A MIXING MODEL DUE TO VAN DE VUSSE BY A SIMPLE PROBABILITY METHOD [J].
GIBILARO, LG ;
KROPHOLLER, HW ;
SPIKINS, DJ .
CHEMICAL ENGINEERING SCIENCE, 1967, 22 (04) :517-+
[6]   AGGREGATION OF MANY COUPLED CONSECUTIVE FIRST-ORDER REACTIONS [J].
GOLIKERI, SV ;
LUSS, D .
CHEMICAL ENGINEERING SCIENCE, 1974, 29 (03) :845-855
[7]  
Iordache O., 1981, Buletinul Institutului Politehnic `Gheorghe Gheorghiu-Dej' Bucuresti, Seria Chimie - Metalurgie, V43, P93
[8]  
IOSIFESCU M, 1969, RANDOM PROCESSES LEA, P198
[9]   LIMIT THEOREM FOR PARTIALLY OBSERVED MARKOV CHAINS [J].
KAIJSER, T .
ANNALS OF PROBABILITY, 1975, 3 (04) :677-696
[10]  
KEMENY JG, 1960, FINITE MARKOV CHAINS, P123