Selecting Candidate Sections

1
Selecting Candidate Sections

• When Funds are Constrained

2
Network-Level Elements

• Inventory
• Condition Assessment
• Determination of needed work funds
• Identification of candidate projects
• Determination of impacts of funding alternatives
• Feedback

3
CANDIDATE SECTIONS VERSUS SELECTED SECTIONS

• Due to limited information available in most
  network-level systems
• Considerable manipulation of sections, treatments
  timing during
• project selection-level
• project-level analysis
• Sections selected at network-level are candidates
  for selection at project selection project level

4
Ranking/Optimization

• Process of identifying the set of potential
  sections that give the best return for the
  funds spent
• Of course, determining how to measure that return
  is the area of quantifying benefits

5
Ranking or Prioritization

• Ranking based on a set of rules
• Those in worst condition repaired first
• Severe cracking repaired first
• Those with highest benefit-cost ratio repaired
  first

6
Ranking Requirements

• Needs analysis completed first
• Identify those that need work based on a set of
  rules
• Rank those needs based on another set of rules
• Fund from highest to lowest rank until funds
  exhausted
• Can be repeated for subsequent years
• Those not selected can be considered in following
  years
• Funding can vary among years

7
Non-Economic Approaches to Ranking

• Damage
• Repair those with greatest damage first
• Performance function
• Repair those with poorest condition index first
• Composite criteria (priority number)
• 0.5FI 0.5EI 0.5RI
  0.5CI
• PN -------------------- --------------------
  -
• Usage
  ADT

8
Worst First Neglects Preservation Return on
Investment
9
Pay Me Now

• 3 Seal Coats at 0.70 /sy - 24 yrs
• 1 Overlay at 3.50 /sy - 8 yrs
• 2 Seal Coats at 0.70 /sy - 16 yrs
• Total 7.00 /sy for 56 yrs

10
Pay Me Later

• 2 Remove Replace at 14.00 /sy
• 54 yrs
• Total 28.00 /sy for 54 yrs

11
Compare

• Pay Me Now
• Total 7.00 /sy for 56 yrs
• Pay Me Later
• Total 28.00 /sy for 54 yrs
• Which Gave Better Service?

12
Economic Based Ranking Systems

• Least first cost
• Least present costs or equivalent uniform annual
  costs
• Greatest benefit/cost ration
• or greatest effectiveness - cost ratio

13
Least First Cost

• Given defined constraints such as
• Alternative must be designed to last at least X
  (say 40) years
• Alternative must provide some minimum level of
  service over that time
• Select alternative with lowest cost

14
Least Present Value (Cost)

• Select treatment with least present value
• All current future costs converted to present
  value or EUAC

PV SFC 1/(1 I)n
PV present value FC cost of future activity i
interest (or discount) rate n number of years
until future activity
15
Least Equivalent Uniform Annual Cost
Annualized present value
where EUAC equivalent uniform annual
cost PV present value i interest (or
discount) rate n number of years in analysis
period
16
Benefit/Cost Ratio

• Net monetary benefits gained by public divided by
  total costs incurred by responsible agency over
  some selected design period

S Benefits to public B/C
------------------------- S Costs to
agency
Benefits costs can be positive and negative
17
Cost-Effectiveness Analysis

• Same as benefit cost - except
• surrogate used for benefit
• Must select fixed cost or fixed effectiveness
• Select measure of effectiveness
• Area under performance curve
• Remaining life
• Normally must develop weighting for usage

18
Prioritization Based on Cost-Effectiveness
19
Cost-Effectiveness

• Sections that will be in the best condition for
  the longest time for least cost
• Give best return on funds
• Should be repaired first

20
Cost-Effectiveness Ratio
Effectiveness
Cost-Effectiveness ---------------------
Cost
21

• Which sections last the longest?
• Which cost the least to build?
• Must generally weight for usage

22
Weighted Cost-Effectiveness Ratio
Area under Curve
Weighted Cost- Effectiveness
------------------------- X WF
EUAC/yd2 (m2)
23
Weighted Cost-Effectiveness Ratio
Where WER weighted effectiveness
ratio AREA area under PCI curve YR
years affected WF weighting factor for
usage EUAC equivalent uniform annual cost
SY square yards in management section
24
Weighting Factor

• Selected for usage
• Vehicle miles traveled
• Volume delivered
• Homes served
• Customers served
• Some function (log, etc.)

25
Start Tues
26
Optimization

• Selects best combination of
• Segments
• Treatment categories
• Application times
• Over selected analysis period

27
Typical Formulation
Maximize or Minimize
Subject to
28
where Xijt section i of m total sections with
alternative j of k total treatment alternatives
in year t of the analysis period Bijt present
value of annual benefits of section i with
alternative j built in year t, all discounted to
a base year at a selected discount rate Dijtt'
the actual treatment cost of section i with
alternative j built in year t, incurred in year
t Bt budget for year t'
29
Optimization Procedures

• Linear programming
• Non-linear programming
• Integer programming
• Dynamic programming
• Heuristics

30
Linear Programming
31
Solution to Linear Programming
32
Non-Linear Programming
33
Integer Programming
34
Dynamic Programming
35
Hueristic Solutions
36
Optimization Process

• Establish objective
• Define benefit (monetary or effectiveness)
• Formulate mathematical model
• Solve mathematical problem

37
Annual versus Long-term Optimization

• Annual
• Selects which sections what treatment
• Repeat for each year
• Long-term
• Selects which sections, what treatment, when to
  apply

38
Annual Optimization
39
Long-Term Optimization
40
Probabilistic (or Fuzzy) versus Deterministic

• Considers inherent variability in future changes
• Typical
• Markov
• Survivor probability

41
Other Approaches

• Artificial Neural Network
• Fuzzy Neural Network
• Genetic Algorithms

42
Multivariate Analysis

• Consider more than one objective
• Goal programming
• Multivariate utility analysis

43
Optimization Limitations

• Work best in limited solution space
• Many infrastructure problems have large solution
  spaces
• Small changes in constraints may result in jumps
  among changing local optimal points

44
Near Optimization(Multiple Year Prioritization)

• Allows identification of
• Sections
• Treatment categories
• Application time??

45
Approaches

• Marginal cost-effectiveness
• Incremental cost-benefit analysis

46
Steps

• 1. Identify the feasible treatments
• 2. Calculate the effectiveness (E)
• 3. Calculate the cost (C) in net present value
  terms
• 4. Calculate the cost-effectiveness (CE) as the
  ratio of E/C
• 5. Select the treatment alternative and time for
  each section with the best CE until the budget is
  exhausted

47

• 6. calculate the marginal cost-effectiveness
  (MCE) of all other strategies for all sections as
  follows
• MCE (Es - Er)/(Cs - Cr)
• where
• Es effectiveness of the strategy selected in
  step 5
• Er effectiveness of the strategy for comparison
• Cs cost of the strategy selected in step 5
• Cr cost of the strategy for comparison

48

• 7. If the MCE is negative, or if Er is less than
  Es, the comparative strategy is eliminated from
  further consideration if not, it replaces the
  strategy selected in 5
• 8. This process is repeated until no further
  selections can be made in any year of the
  analysis period

49
Factors Affecting Network-Level
Prioritization/Optimization
50
Management Organization
51
(No Transcript)
52
(No Transcript)
53
(No Transcript)
54
Multiple Funding Sources

• Especially
• Non-exclusive funds
• Interdependent funding requirements

55
(No Transcript)
56
Prioritize After Fund Allocated to Funding
Category
57
(No Transcript)
58
(No Transcript)
59
(No Transcript)
60
Resources of the Agency

• To conduct process
• To fund treatments

61
Reality of Use

• How will results be applied?
• Candidate section versus selecting sections and
  treatments
• What will changes do to credibility?
• How will list be changed with small changes in
  resources