@inproceedings{oai:kutarr.kochi-tech.ac.jp:00001241,
 author = {MIZUTANI, Daijiro and KAITO, Kiyoyuki and KOBAYASHI, Kiyoshi},
 book = {Society for Social Management Systems Internet Journal},
 month = {Dec},
 note = {It is necessary to identify infrastructures relatively deteriorating fast, and to monitor, repair and
renew the infrastructures. However, regarding methods to extract the infrastructures from inspection data
obtained through ordinary inspections, there is actually no systematized methodology. In this paper, the
authors propose the multi-stage mixed Markov deterioration hazard model and its multi-hierarchical
Bayesian estimation. Furthermore, the benchmarking analysis towards stratified deterioration speeds
corresponding to decision making levels and the methodology to extract intensively monitored
infrastructures on each level are proposed. In order to verify the effectiveness of the proposed methodology,
empirical analysis is carried out using the visual inspection data of 10,689 expansion joints in 21 lines of
actual highway. The authors first mentioned that the deterioration process significantly depends on (1) kind
of expansion joint, (2) kind of surface layer pavement and (3) traffic volume, and clarified that the expected
life span of the expansion joints is about 18 years and it varies about 5 years due to the above mentioned
factors. Then, it was found that the expected life span of the expansion joints varies from about 9 years to
about 55 years by considering the heterogeneity of each line. Furthermore, the authors clarified that the
expect life span varies from about 5 years to more than 100 years in the fastest deteriorating line by
considering the heterogeneity of each expansion joint. Finally, by using the estimated result, the authors
carried out the relative evaluation of hazard rate and extract the intensively monitored expansion joints.},
 publisher = {Society for Social Management Systems},
 title = {EXTRACTION OF INTENSIVELY MONITORED EXPANSION JOINTS BY MULTI-STAGE MIXED MARKOV HAZARD MODEL},
 volume = {9},
 year = {2014}
}