Incremental Costs and Performance Benefits of Various Features of Concrete Pavements

Previous | Table of Contents | Next

CHAPTER 3. ANALYSIS METHOD

Introduction

As described in, the analysis approach used in this project is a simple method to investigate the cost and performance implications of changing design features in a PCC pavement. The primary purpose of the analysis method is to compare two different defined pavement sections (i.e., two sections with different design features) to determine the relative differences in their expected costs and expected performance (in terms of allowable ESALs), Although the calculations used in the methodology are simple and straightforward, many components are involved in the overall process. This chapter defines the basic components of the methodology, provides illustrative examples demonstrating the computational methods of the approach, and introduces the analytical software that automates the methodology.

Components of the Analysis Approach

Before discussing the detailed performance and cost computations, it is important to first explain the general components making up the approach. Specifically, this section discusses the definition of pavement sections (and in particular the Standard pavement section), cost and performance data sets, category ranking factors, and the simplified life-cycle cost analysis (LCCA) included as part of an analysis session.

Definition of a Pavement Section

A pavement section is defined as a unique combination of specific pavement features chosen from the following 10 different design feature categories:

• Subgrade.
• Base/Subbase.
• Drainage.
• Thickness/Slab Size.
• Cross Section.
• Joints/Load Transfer.
• Joint Sealing.
• Shoulders.
• Strength/Materials.
• Initial Smoothness.

For practical purposes, the number of available design feature options had to be limited to those alternative design features (organized by design category) previously outlined in the Design Categories and Alternative Design Features section of chapter 2.

The Standard Pavement Section

Because the goal of the survey of user agencies and contractors was to estimate the expected changes in cost and performance associated with making one design feature change at a time, a Standard pavement design was defined to serve as a baseline. The unique combination of design features for the Standard section are listed in table 7. This defined Standard pavement section is expected to carry approximately 700,000 to 800,000 ESALs per year over an assumed 20-year design life (design life is defined as the time until first major rehabilitation).

As part of the survey, the respondents were asked to use the expected cost and performance of the Standard section as a basis while they systematically replaced one of the Standard section's features with one of the other available alternative design features. With every feature replacement, respondents were asked to estimate expected percent changes in terms of cost and performance ratios. That is, the cost and performance ratios for the Standard pavement section were assumed to be 1.0. If a particular design feature change was expected to result in a 5-percent increase in cost and a 2-percent decrease in performance, then the survey respondent would have reported cost and performance ratios of 1.05 and 0.98, respectively.

Cost and Performance Data

The most basic part of the methodology is based on estimating the total change in cost and performance associated with changing one or more design features from the Standard pavement section. Within the analysis tool, these relative percent changes in cost and performance are summarized into cost and performance data sets. Each data set is defined as a summary of the relative percent changes in cost or performance associated with all available design feature values in each of the 10 design categories (i.e., all of the available design features outlined in chapter 2). The following sections contain specific discussions of how the summarized cost and performance data are used to estimate an overall section cost and performance within the analysis procedure.

Relative Cost Data

When multiple design features are changed from the Standard section, the percent change in costs of each feature can be simply summed to determine the cost of the modified section (i.e., cost changes are cumulative). This procedure is believed to be valid because most of the costs of additional design features are independent construction variables. For example, assume that a pavement section differs from the Standard pavement section in that the 150-mm (6-inch) dense-graded aggregate base is replaced with a 150-mm (6-inch) cement-treated base, and the hot-poured asphalt with widening cut joint seal is replaced with a preformed compression sealant. Next, assume that the base type and joint sealant changes are expected to result in 6.0 and 7.0 percent respective cost increases. The construction of these two design features is completely independent, so these costs can be directly added for a total expected cost increase of 13.0 percent for this pavement section.

Relative Performance Data

Unlike cost data, the impact of individual design features on performance is not necessarily cumulative because most, if not all, design variables are interdependent. This means that performance impacts are also different depending on what other design features are present in the pavement section. For example, tied PCC shoulders might have a larger impact on performance if the pavement is constructed with an aggregate base rather than a stabilized base. Similarly, the effect of a permeable base may be greater if the joints are not doweled. To address this challenge, the underlying methodology relies on a simple method to calculate the impact of performance when multiple design features are changed simultaneously. This method uses a ranking system that rates the design features based on impact and importance. These impact rankings, referred to as Category Ranking Factors, are discussed in greater detail in the next section.

Category Ranking Factors

One of the most difficult steps in the analysis approach is the assignment of category ranking factors that accurately reflect an agency's assessment of which design feature categories have the largest impact on overall performance. Because the assignment of realistic ranking factors is imperative for computing realistic overall performance values, a detailed discussion illustrating the importance of these ranking factors is helpful to explain the meaning of ranking factors, illustrate their use, and identify possible pitfalls to avoid.

Background

For those cases when multiple design features change at the same time, a method is needed to estimate the overall expected performance change. Historically, many different methods of different complexity levels have been used to make this performance estimate. By far the most complex approach is to conduct extensive research to develop mechanistic performance models as functions of the changing design features. While this most complex approach would likely provide the most accurate results, it is obviously beyond the scope of this project. More simple methods that have been used include estimating overall performance as the product, sum, average, or weighted average of the different individual performance changes associated with each changing design feature.

For this project it was decided to adopt a method in which the relative importance of the 10 different design categories would be used to weight each individual performance change. These relative weighting factors are referred to as category ranking factors. More specifically, ranking factors are used to determine how much of each individual performance value is added to the overall modified performance.

Obtaining more accurate estimates of overall performance would have required including the entire spectrum of design feature combinations in the questionnaire surveys, or the use of reliable performance prediction models that incorporate every design feature. Because these approaches were not possible under this project, the described approach was adopted as a simple and expedient measure of estimating overall performance.

Practical Range for Overall Performance 

Before discussing the detailed computations associated with applying category ranking factors, it is helpful to conduct a reasonableness check by estimating the practical range for the overall expected change in performance. For the category ranking factor method adopted in this analysis approach, the following may be used to define practical limits that establish a range of reasonable values:

• If all individual percent changes in performance are positive, the overall modified value for all practical purposes lies between the smallest individual value and the sum of all individual values.
• If all individual percent changes in performance are negative, the overall modified value most likely lies between the greatest individual value (i.e., that closest to zero) and the sum of all individual values.
• If the individual percent changes in performance contain both positive and negative values, the overall modified value most likely lies between the sum of the negative performance ratings on the lower end, and the sum of the positive performance ratings on the upper end.

To demonstrate these practical limits with an example, assume that relative performance values of +8.0 percent, -1.0 percent, and +15.0 percent are associated with changes in Subgrade, Base/Subbase, and Drainage, respectively. Because this example contains both positive and negative performance changes, the third practical limit situation applies. Therefore, the modified performance value should be expected to be between -1.0 and +23.0 percent. The application of category ranking factors will always result in a modified performance value that lies within these established practical limits.

One might argue that the largest individual performance change would define the minimum expected overall performance for every analysis case. For example, if individual performance values of +6.0 percent, +2.0 percent, and +12.0 percent are associated with changes in Subgrade, Base/Subbase, and Drainage, one might conclude that the overall performance would be between +12.0 percent (the largest individual value) and +20.0 percent (the sum of all values). While this approach may hold true for some cases, without understanding all of the interdependencies among individual design features, it is more conservative to describe the practical limits as defined in the bulleted list above.

Category Ranking Factors Sets

A category ranking factor set is a prioritized list of the 10 different design categories; it reflects their relative importance on the overall performance of a PCC pavement. A ranking factor set is defined by first sorting all 10 design feature categories in order of decreasing importance. Next, integer values from 10 to 1 are assigned to the sorted category list (10 is assigned to the category deemed most important and 1 to the least important category). For the analysis procedure to work as intended, no two design categories are allowed to be assigned the same integer ranking value. To form a basis for this process, SHA survey respondents were asked to submit their assumed category ranking set along with their complete performance data set. The survey results were then used to prepare the default category ranking factor set presented in table 5. It is emphasized that the resulting overall performance change predicted using category ranking factors is merely an estimate of the true performance change that would be realized.

Application of Category Ranking Factors

Category ranking factors are defined as relative because for a given investigation, normalized ranking multipliers are computed by dividing each individual ranking factor by the largest of the included ranking factors (i.e., for only those design features changing from the Standard pavement section). This method can be demonstrated by using the example displayed in table 8, where relative performance values of +1.0 percent, +7.0 percent, and +10.0 percent are associated with changes in Subgrade, Base/Subbase, and Drainage. Next, assume individual ranking factors of 5, 8, and 7 are associated with the three respective design feature categories. The largest impact factor of the three included feature categories is the "8" associated with Base/Subbase. Therefore, all three of the included impact factors are divided by "8" to compute normalized ranking multipliers. These normalized ranking multipliers are then multiplied by the associated expected relative performance values to give a modified performance value for each design category. The overall section performance is then determined as the sum of all modified performance values. For this example, the expected increase in performance is estimated to be 16.4 percent.

In this example, because Base/Subbase is assigned the largest ranking factor, by definition its whole associated individual performance is added to the overall modified performance value (i.e., its individual performance change is deemed the most important of the three associated individual performance values). That is, the +7.0 percent expected relative performance associated with Base/Subbase is the starting point for the overall performance computations. Next, the individual performance changes associated with the other included design feature categories (Subgrade and Drainage) are, therefore, diminished by dividing each associated ranking factor by the largest included ranking factor.

It is important to note that within this methodology, the relative differences between ranking factors (rather than the actual ranking factor values) are important when determining overall modified performance. For example, one might think that the design feature assigned an impact factor of 10 is always going to be important when determining the overall pavement section performance. The previous example shows that this is not the case, as none of the three changing design feature categories had a category ranking factor of 10. Normalizing all individual ranking factors to the largest of the included factors ensures that the performance of the most important included design feature becomes the starting point of the modified performance computation. In the example, it is noted that if Base/Subbase were the only design feature category that was changing, then the total modified performance would be +7.0 percent. Therefore, the other design features deemed less important are, in a sense, used to adjust the +7.0-percent value associated with Base/Subbase. The normalized ranking multipliers give an indication of the relative impact of the adjustments.

Selection of Appropriate Ranking Factors

As mentioned previously, the assignment of realistic category ranking factors is one of the most important and challenging steps of the analysis approach. As with any methodology that has built-in functionality, it is possible to misuse this method and achieve results that are counter-intuitive. This section is intended to provide guidance on selecting ranking factors so as to avoid such pitfalls.

The first recommendation is that it is very important to assign category ranking factors that do not contradict the performance values observed within different design categories. That is, those design categories where the largest percent increases or decreases in performance are observed should most likely be the design categories with the largest category ranking factors. For example, assume that investigated Thickness/Slab Size choices result in individual performance changes from -40.0 to +50.0 percent, while different Joint Sealing choices result in a range of individual performance between -5.0 and +5.0 percent. For this case, the category ranking factor assigned to Thickness/Slab Size should be significantly larger than that assigned to the Joint Sealing. To illustrate this point, the two examples (although extreme) shown in tables 9 and 10 illustrate how important it is to assign category ranking factors that reflect the largest impacts on overall performance.

Note that in table 9 the category ranking factors are intuitively assigned to reflect the relative importance of the two design feature categories; the total modified performance value appears reasonable at +45 percent. Table 10 illustrates the same example but with the category ranking factors reversed. For this case, the overall modified performance is computed to be 0.0 percent, which appears to be a counterintuitive value as one would expect the larger percent performance change associated with the Thickness/Slab Size to control. While the analysis method and associated software will allow the user to set these category ranking factors to any number, it is very important to understand how they will be used in the computations.

It is equally important to remember that the modified performance values resulting from the application of ranking factors are additive. The individual performance associated with the largest ranking factor is used as the starting point (e.g., 7.0 percent was used as the starting point of the overall performance computation in table 8). This is because the category with the largest associated ranking factor is assumed to have the largest influence on overall performance. That is, in many cases, one should expect the overall performance to be close to the one individual performance value associated with the largest ranking factor as it, by definition, is the governing performance value. As mentioned previously, the ranking factors associated with other included design features are used to diminish those associated individual performance changes before adding them to the overall performance calculation. That is, all design feature categories that are not deemed to be the most important category (i.e., their ranking factors are less than the largest included ranking factor) are simply used to adjust the individual performance change associated with the largest ranking factor. It is the defined ranking factors that are used to determine the ranking factor ratios (normalized ranking multipliers) that determine the how much of each individual performance change is added to the overall performance value.

Another way to assess the relative meaning of these impact factors is to compare the included values directly. For instance, in the example presented in table 8, based on the entered Subgrade, Base/Subbase, and Drainage ranking factors of 5, 8, and 7, respectively, one can deduce that the user assumed the following:

• The individual performance value associated with Subgrade is 5/8 as important as the individual performance associated with Base/Subbase.
• The individual performance value associated with Drainage is 7/8 as important as the individual performance associated with Base/Subbase.
• The individual performance value associated with Subgrade is 5/7 as important as the individual performance associated with Drainage.

An earlier section of this chapter (Demonstration of the Analysis Method) illustrates how ranking factors are used in the computations. This should help emphasize the importance of selecting appropriate impact factors that correspond with the assumed impact on performance associated with each design category. The selection of user-defined category ranking factor sets should be carefully considered so that appropriate values reflective of observed performance are developed.

Simplistic Life-Cycle Cost Analysis

Because design feature changes alter the expected performance (estimated service life) of a given pavement section, the associated life-cycle cost (LCC) stream is also affected. To investigate the magnitude of the impact of design feature changes on LCCs, the analytical tool does provide a means of conducting a simplistic LCCA as part of the analysis. However, because of its simplistic nature, the user of the software tool is warned that the results of the LCCA results should be viewed with caution. While the cost trends may be realistic, the actual computed dollar values may or may not be accurate. If more accurate LCCA results are desired, it is recommended that a more rigorous LCCA be conducted using established methods.⁽¹⁰⁾

The LCCA conducted within this software is described as simplistic in that the cost stream values (annual maintenance, rehabilitation, and salvage value costs) can all be determined using simplified methods. Specifically, the following important components are included in the simplistic LCC approach:

• Standard Pavement Section Design Life-The expected design life (in years) of the defined Standard pavement section. Based on the design features used to define the Standard pavement section, this design life is assumed to be 20 years (note: this 20-year design life is hard coded in the software). Other pavement section design lives are computed by multiplying this 20-year design life by the expected performance ratio computed by using the analysis methodology. For example, if a custom section was found to have an overall modified expected performance of +7.0 percent, the expected service life of the custom section is computed as 1.07 * 20 years = 21.4 years.
• Initial Construction Cost for the Standard Pavement Section-Fence-line-to-fence-line pavement-related cost expressed in dollars per lane-mile.
• Analysis Period-Period of time over which future maintenance and rehabilitation costs are to be considered in the LCCA. Based on FHWA recommendations, the analysis period must be a minimum of 35 years.
• Discount Rate-The discount rate, entered as a percentage, is the estimated difference between the interest and inflation rates over a long time period. It is used to translate actual LCCs into equivalent present worth costs. Historically, the discount rate has been in the range of 3 to 5 percent.
• Salvage Value-The user has the choice whether or not to include a salvage value in the LCCA. If it is included, a linear deprecation of the rehabilitation cost is used to determine the salvage value at the end of the analysis period. For example, if a rehabilitation with an expected life of 10 years and a cost of $50,000 is applied 3 years before the end of the analysis period, the salvage value would be computed as $50,000 * (10 - 3)/10 = $35,000. The salvage value is essentially a negative cost applied at the end of the analysis period.
• Time Until First Rehabilitation-The time until first rehabilitation may either be defined as a percentage of the expected design life or as an actual pavement age (in years). This input is used to determine the time at which the rehabilitation cost is to be included in the LCCA. For example, if the design life is 20 years, and the time until first rehabilitation is entered as 80 percent, then the rehabilitation cost included in the LCC stream would be applied at a time of 20 years * 0.8 = 16 years.
• Assumed Rehabilitation Life-The assumed life (in years) of the chosen rehabilitation activity. As subsequent rehabilitation activities are assumed to be applied until the end of the analysis period, this LCC input determines the years in which those subsequent rehabilitation activities will be applied.
• Cost of Annual Maintenance-The cost of annual maintenance may either be entered as a percentage of the provided initial construction cost or as an actual cost. This input is used to determine the maintenance cost applied annually within the LCCA. For example, if the initial construction cost is $300,000 per lane-mile, and the cost of annual maintenance input is set to 5 percent, then the annual maintenance costs included in the LCC stream are computed as 0.05 * $300,000 = $15,000/lane-mile per year.
• Rehabilitation Cost-The cost of rehabilitation may either be entered as a percentage of the provided initial construction cost or as an actual cost. This is used to determine the cost of the rehabilitation activity that is to be applied at regular intervals (after first application) until the end of the analysis period.

It should be noted that all annual maintenance and rehabilitation-related LCCA inputs are section-specific within the analysis approach. That is, all of these LCC inputs may be customized for each unique pavement section that is defined. The primary purpose of linking these cost inputs to a section is to accommodate the many cases where the inclusion of a design feature directly influences the future maintenance and rehabilitation costs associated with that section (e.g., including edge drains will result in the additional cost of cleaning the edge drains).

The LCCA for a given pavement section is dependent on the expected cost and performance ratios computed for that section. Specifically, as a first step, these ratios are used to compute the expected initial cost (in actual dollars) and expected design life (in actual years) used to compile an associated cost stream. As an example, assume that cost and performance ratios of 1.07 and 1.05, respectively, are computed for a given pavement section. In addition, assume a design life of 20 years and an initial construction cost of $500,000 for the Standard pavement section. Therefore, the custom pavement section is assumed to have an expected service life of 20 years * 1.05 = 21 years, and an initial construction cost of $500,000 * 1.07 = $535,000. These values are used as the basis of the LCCA associated with the custom pavement section. For a given pavement section, the primary outputs of the simplified LCCA are the individual computed costs (annual maintenance, rehabilitation, and salvage value), their total present worth values, the summarized equivalent uniform annual cost (EUAC), and ratios of the EUACs between the different sections being compared.

Demonstration of the Analysis Method

As previously discussed, the underlying analysis method, in its most basic form, is based on making cost and performance comparisons between a user-defined pavement section and the Standard pavement section. This underlying methodology can be expanded so that two user-defined sections can be compared to each other after each has been individually compared to the Standard pavement section. The computation methodology used to make such comparisons is described below.

Comparing a Custom Pavement Section to the Standard Pavement Section

The following example illustrates the computations involved when comparing a user-defined (modified) pavement section to the Standard section. In this example, the modified pavement section is defined by changing three different design features from the Standard section. The specific feature changes are summarized in table 11. The expected cost and performance changes and applicable category rankings for the example are defined in table 12.

As mentioned previously, the costs associated with different design feature changes are cumulative. Therefore, for this example (table 12), the change in cost is computed to be +16.0 - 1.0 + 7.0 = +22.0 percent (i.e., the modified section is expected to cost 22.0 percent more than the Standard pavement section).

Because the performance changes include both positive and negative values, the practical range of performance is estimated to be between the sum of the negative values on the lower end and the sum of the positive values on the upper end. Therefore, for this example the overall performance change is expected to be between -5.0 percent (the one negative value) and +10.0 percent (the sum of the two positive values). The category ranking factors are used to estimate the weighted contributions of each individual performance rating. The details of this computation process are illustrated in table 13.

Normalized ranking factor multipliers are determined for each changing design feature category by dividing each individual category ranking factor by the largest of the included category ranking factor value. For this example, the largest included category ranking factor is the 10 associated with the Joints/Load Transfer design feature category. Modified performance values are computed by multiplying the individual expected relative performance values by the computed normalized ranking multipliers. Summing the individual modified performance ratings for this example gives a total estimated increase in performance of 0.5 percent.

Comparison of Two Custom Pavement Sections (Section A vs. Section B)

Although the previous example compared a defined custom pavement section to the Standard pavement section, the methodology also allows the user to compare one custom section to another (e.g., Section A to Section B). The first step of this process is to make the following two independent comparisons: 1) compare Section A to the Standard pavement section and obtain expected percent cost and performance changes, and 2) compare Section B to the Standard pavement section and obtain expected percent cost and performance changes. The final step of the process is to compare the individual results so that the expected cost and performance of Section B can be reported in terms of the expected Section A cost and performance. The following example demonstrates this process.

Step 1: Compare Section A to the Standard Pavement Section

Assume that the goal of this example is to compare the two pavement sections summarized in table 14 (note that the shaded cells in table 14 indicate that those design features do not differ from the Standard pavement section). As indicated, the first step is to compare each individual pavement section to the Standard pavement section. Noting that Section A is the pavement section used in the previous example, its analysis details were those summarized in table 13. Therefore, comparing Section A to the Standard pavement section found that the expected cost and performance change expected with Section A are the following:

• Expected relative cost change: +22.0 percent.
• Expected relative performance change: +0.5 percent.

Step 2: Compare Section B to the Standard Pavement Section

A closer look at Section B finds that only Base/Subbase type and Strength/Materials differ from the Standard pavement section. The analysis details for Section B are summarized in table 15.

Since the costs associated with different design feature changes are cumulative for this example, the change in cost +6.0 + 5.0 = +11.0 percent (i.e., Section B is expected to cost 11.0 percent more than the Standard pavement section).

The true measure of performance for Section B is assumed to be somewhere between the lowest individual performance value and the sum of all of the values. Therefore, for this example the performance rating is between +5.0 percent (the lowest individual value) and +15.0 percent (the sum of all of the values). The category ranking factors are used to estimate the weighted contributions of each individual performance rating. This computation process is illustrated in table 16.

Step 3: Comparison of Sections A and B

Finally, a simple comparison of the expected relative cost and performance changes for both Sections A and B is made by reporting the Section B cost and performance as ratios of the expected Section A values. Table 17 summarizes the expected relative cost and performance changes from the individual comparisons with the Standard pavement section.

Therefore, a ratio of expected Section B costs to expected Section A costs is computed as (1.11/1.22) = 0.91 (i.e., the cost of Section B is expected to cost 0.91 times the cost of Section A). Using the same approach, a ratio of expected Section B performance to expected Section A performance is computed as (1.125/1.005) = 1.12 (i.e., Section B is expected to carry 1.12 times as many ESALs as Section A).

Step 4: Interpreting Benefit/Cost Ratios

Other interesting information can be gleaned by looking at the benefit/cost (B/C) ratios computed for each pavement section. These B/C ratios are informative in that they allow an assessment of the cost-effectiveness of a defined combination of different design features. In a comparison of two pavement sections, the section with the largest B/C ratio is the most cost effective section to construct. Another way to interpret this B/C ratio is that the larger the B/C ratio, the more performance achieved per dollar spent.

In the example above, the Section A numbers indicate that the chosen set of design features will cost 22 percent more, but will only result in a 0.5 percent increase in performance. These cost and performance numbers translate to a B/C ratio of 0.82. In contrast, the Section B numbers indicate that the chosen set of design features will cost 11 percent more, but will result in a 12.5 percent increase in performance. These Section B numbers translate to a B/C ratio of 1.01. Therefore, a comparison of B/C ratios indicates that Section B is more cost effective to construct than Section A (1.01 is greater than 0.82).

Introduction to the Analytical Software Tool

A major part of this project is the development of an analytical software tool that automates the analysis method described in the subsequent sections of this chapter. The software is provided as a tool for pavement engineers or contractors who are interested in investigating estimated cost versus performance trade-offs associated with the selection of different design features during the concrete pavement design process; it is absolutely not intended as a "design" tool. Instead, it provides a "reasonableness" check regarding the "justification" or "questioning" of the addition of different design features. Also, because the default cost and performance changes in the software are estimated based on collected survey data from all over the United States, it is strongly suggested that the user define cost sets, performance sets, and category ranking factor sets that reflect local experiences and conditions. That is, the results from this tool are only as good as the data on which they are based. Finally, it is again emphasized that the output results from this tool are solely "estimates" of cost and performance associated with changing design features and, therefore, should be used with caution.

The software utilizes a "modular" data storage approach in that all unique user-defined pavement sections, cost sets, performance sets, and category ranking factors sets are defined, named, and saved to respective "master lists" within the software. Therefore, the process of defining an analysis session is simplified in that the user simply builds an analysis session by selecting these named and saved data groups from their respective master lists. This approach was adopted because of the large number of data elements that need to be defined when conducting an analysis. When getting started with the software, users are encouraged to define all of the modular data groups anticipated for use in the analysis sessions. Much more detailed descriptions of the software's components are contained in the Software User's Guide included as appendix D of this report.

Once all unique pavement sections, cost data sets, performance data sets, and category ranking factor sets of interest are defined and saved within the software, the user may then focus on building specific analysis sessions of interest. Specifically, two general types of analysis sessions may be conducted using the software:

1)  Direct Comparison-The Direct Comparison analysis is used to compare two defined pavement sections to assess expected differences in cost and performance. A by-product of this type of analysis is the B/C ratio associated with each section. In a comparison of two pavement sections, the section with the largest B/C ratio is the most cost effective section to construct (the larger the B/C ratio, the more performance achieved per dollar spent).

• Sensitivity Analysis-The Sensitivity Analysis is provided as a method of defining more complex relationships among pavement designs. Specifically, the following two general types of sensitivity analyses may be defined in the software:
• Comparison of multiple pavement sections-The first sensitivity analysis type that may be defined is the comparison of one pavement section to a number of other pavement sections with different design features. This series of direct section comparisons is conducted while holding all other analysis inputs constant.
• Sensitivity of a chosen pavement section to other process inputs-The second sensitivity analysis type allows the user to investigate the sensitivity of one defined pavement section to changes in the other inputs of the process. Specifically, these sensitivity analyses are used to compute cost and performance changes associated with subjecting a chosen pavement section to a series of different cost data sets, performance data sets, category ranking sets, or a combination of both cost and performance sets. Each of these different data sets is discussed in detail in the Software User's Guide included as appendix D.

The results of any defined analysis session are summarized in a customizable output report that may be previewed in an on-screen window or printed. The detailed output report contains all details of the analysis session including a list of input values, tables of intermediate cost and performance computations, and a series of summary tables for cost, performance, and LCCA results.

Summary

This chapter outlines the simple analysis approach used in this project to estimate overall cost and performance changes associated with changing pavement design features. The basis components of the analysis procedure are first discussed in detail to explain the underlying theory used to estimate overall changes in cost and performance. Next, the analysis method is demonstrated through examples to further clarify the details of the analysis method. Finally, the associated analytical software tool is introduced with specific discussion of the intended usage and limitations of the software. A detailed explanation of how to define and conduct analysis sessions using the analytical software tool is contained in the Software User's Guide (included as appendix D).