PI: Jes

1
CMMI- Information Technology and Infrastructure
Systems
Grant CMMI-0726789 A Comprehensive Framework
for the Efficiency Measurement of Road Maintenance
PI Jesús M. de la Garza (chema_at_vt.edu), Virginia
Tech Co-PI Konstantinos Triantis
(triantis_at_vt.edu), Virginia Tech SP Mehmet E.
Ozbek (meozbek_at_vt.edu), Virginia Tech

• 1988 A survey performed on about 10 of all USA
  infrastructure by the National Council on Public
  Works Improvement revealed that the nations
  roads were in better than fair condition (Mirza
  2006).
• 1998
• 2001
• 2003
• 2005
• The Federal Highway Administration endorsed
  asset management to be the future approach of
  road maintenance for all state departments
• of transportation (DOTs) (JLARC 2002).

Current Road Maintenance Performance Measurement
Systems
Similar surveys by American Society of Civil
Engineers revealed that the nations roads were
in poor condition (Mirza 2006).

• Do not investigate the effect of the
  environmental factors, e.g., climate and
  location.
• Do not investigate the effect of the operational
  factors, e.g., traffic, load, design-construction
  adequacy.

• Solely focus on effectiveness measures, e.g.,
  level-of-service.
• Disregard the efficiency concept, e.g., the
  amount of resources utilized to achieve such
  level-of-service.

• In

Not knowing how efficient state DOTs are in
being effective can lead to excessive and
unrealistic maintenance budget expectations.
For the cases in which comparative analyses
are made, disregarding such external and
uncontrollable factors and using pure
effectiveness results may lead to unfair
comparisons.

• Allocating resources to preserve, operate, and
  manage the nations transportation
  infrastructure.
• Calls for the utilization of management,
  engineering, and economic principles to help
  state departments of transportation (state DOTs)
  in making decisions as to how resources should
  be allocated.
• Requires, as an integral part, performance
  measurement.
• 
  (Geiger 2005)

The findings of the research outlined herein will
contribute new knowledge to the asset management
field in the road maintenance domain by providing
a framework that is able to differentiate
effective and efficient maintenance strategies
from effective and inefficient ones as such, the
impact of such framework will be broad,
significant, and relevant to all transportation
agencies.
1
2
3

• The methodology calls for the development of
  modules to accommodate different scenarios that
  relate to
• the different units of comparison
• availability of data in different degrees
• The specific activities that need to be
  undertaken as a part of the methodology are
• Activity 1 Explore the approaches that can be
  used to deal with the uncontrollable factors. The
  main approaches to be investigated are
• single-stage approaches
• multi-stage approaches (e.g., bootstrapping,
  clustering, and regression)
• Activity 2 Explore the approaches that can be
  used to refine the comprehensive sets of
  controllable variables and uncontrollable
  factors. The main approaches to be investigated
  are
• judgmental approaches (e.g., analytical
  hierarchy process)
• quantitative approaches (e.g., principal
  component analysis)
• DEA-based approaches (e.g., DEA-based
  sensitivity analysis)
• Activity 3 Explore the types of the DEA models
  to be run. The DEA models to be investigated are
  
• input-oriented CCR model
• output-oriented CCR model
• input-oriented BCC model

When there are multiple inputs/outputs
Data Envelopment Analysis (DEA)
Partial Efficiency Measure Approach Has a
potential to result in serious misunderstandings
about the overall efficiency of a process when
only a single partial efficiency ratio is used
(Craig and Harris 1973). Total Factor Efficiency
Measure Approach May result in subjectivity as
the decision-maker prescribes weights to be
assigned to each input and output variable
(Cooper et al. 1999). System Dynamics Requires
the definition of the of mathematical
relationships between key variables (Chasey et
al. 1997). Regression Analysis Compares the
efficiency of units against a hypothetical
average performance (Sexton 1986).

• DEA can simultaneously deal with multiple outputs
  and multiple inputs.
• DEA does not require the specification of a
  priori weights for the variables.
• DEA is non-parametric.
• DEA focuses on the best-practice frontiers.

(Ozbek 2007)
4
6
5
(Charnes et al. 1994, Rouse 1997, Ramanathan
2003)