Detail publikace
Statistical Evaluation of Large-Scale Data Logistics System
ŠOMPLÁK, R. ŠMÍDOVÁ, Z. SMEJKALOVÁ, V. NEVRLÝ, V.
Anglický název
Statistical Evaluation of Large-Scale Data Logistics System
Typ
článek v časopise ve Scopus, Jsc
Jazyk
en
Originální abstrakt
Data recording is struggling with the occurrence of errors, which worsen the accuracy of follow-up calculations. Achievement of satisfactory results requires the data processing to eliminate the influence of errors. This paper applies a data reconciliation technique for mining of data from ecording movement vehicles. The database collects information about the start and end point of the route (GPS coordinates) and total duration. The presented methodology smooths available data and allows to obtain an estimation of transportation time through individual parts of the entire recorded route. This process allows obtaining valuable information which can be used for further transportation planning. First, the proposed mathematical model is tested on simplifled example. The real data application requires necessary preprocessing within which anticipated routes are designed. Thus, the database is supplemented with information on the probable speed of the vehicle. The mathematical model is based on weighted least squares data reconciliation which is organized iteratively. Due to the time-consuming calculation, the linearised model is computed to initialize the values for a complex model. The attention is also paid to the weight setting. The weighing system is designed to reflect the quality of specific data and the dependence on the frequency of trafic. In this respect, the model is not strict, which leaves the possibility to adapt to the current data. The case study focuses on the GPS data of shipping vehicles in the particular city in the Czech Republic with several types of roads.
Anglický abstrakt
Data recording is struggling with the occurrence of errors, which worsen the accuracy of follow-up calculations. Achievement of satisfactory results requires the data processing to eliminate the influence of errors. This paper applies a data reconciliation technique for mining of data from ecording movement vehicles. The database collects information about the start and end point of the route (GPS coordinates) and total duration. The presented methodology smooths available data and allows to obtain an estimation of transportation time through individual parts of the entire recorded route. This process allows obtaining valuable information which can be used for further transportation planning. First, the proposed mathematical model is tested on simplifled example. The real data application requires necessary preprocessing within which anticipated routes are designed. Thus, the database is supplemented with information on the probable speed of the vehicle. The mathematical model is based on weighted least squares data reconciliation which is organized iteratively. Due to the time-consuming calculation, the linearised model is computed to initialize the values for a complex model. The attention is also paid to the weight setting. The weighing system is designed to reflect the quality of specific data and the dependence on the frequency of trafic. In this respect, the model is not strict, which leaves the possibility to adapt to the current data. The case study focuses on the GPS data of shipping vehicles in the particular city in the Czech Republic with several types of roads.
Klíčová slova anglicky
data reconciliation; GPS data; random errors; weight of data; transportation time
Vydáno
21.12.2018
Nakladatel
Brno University of Technology
Místo
Brno, Czech Republic
ISSN
1803-3814
Ročník
24
Číslo
2
Strany od–do
9–16
Počet stran
8
BIBTEX
@article{BUT156433,
author="Radovan {Šomplák} and Zlata {Šmídová} and Veronika {Smejkalová} and Vlastimír {Nevrlý},
title="Statistical Evaluation of Large-Scale Data Logistics System",
year="2018",
volume="24",
number="2",
month="December",
pages="9--16",
publisher="Brno University of Technology",
address="Brno, Czech Republic",
issn="1803-3814"
}