DocumentCode :
1925411
Title :
Sequential observation and selection with value-based payoffs: An experiment study
Author :
Wang Yuan ; Zhong Tong ; Xu Yi-feng
Author_Institution :
Sch. of Manage., Xi´an JiaoTong Univ., Xi´an, China
fYear :
2010
fDate :
8-10 Aug. 2010
Firstpage :
394
Lastpage :
397
Abstract :
Secretary Problem (SP) is labeled as a decision task of sequential observation and selection, which describes the dynamic decision process where relevant information of alternatives is available sequentially, and the essential decision task is whether to choose the current alternative and stop searching further information or to abort the current alternative and wait for the next one. Prior studies focused on how many alternatives the decision maker should observe to maximize the probability of selecting the optimal alternative under the full rationality assumptions, without considering the effects of noise information on decision-making behavior. In this paper, two types of the alternatives´ noise information, variance of the alternatives´ value (VOAV) and the gap of the conjoint alternative´s value (GOCAV) were proposed to affect the number of observations under bounded rationality assumptions, and these was tested by laboratory experiments. The results of our experiments reveal that there is a negative correlation between VOVA and the number of observations, but a positive correlation between the GOCAV and the intention to accept. In addition, the decision-makers often overestimate their ability to target the best with obvious overconfidence bias.
Keywords :
decision theory; dynamic decision process; gap of the conjoint alternatives value; secretary problem; sequential observation; value based payoffs; variance of the alternatives value; Biological system modeling; Context; Context modeling; Educational institutions; Sequential search; experiment study; noise information; number of observation; optimal stopping;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Emergency Management and Management Sciences (ICEMMS), 2010 IEEE International Conference on
Conference_Location :
Beijing
Print_ISBN :
978-1-4244-6064-9
Type :
conf
DOI :
10.1109/ICEMMS.2010.5563419
Filename :
5563419
Link To Document :
بازگشت