Multi-Year Program under Budget Constraints Using Multi-Criteria Analysis

Road investment appraisal requires joint consideration of multiple criteria which are related to engineering, economic, social and environmental impacts. The investment consideration could be based on the economic analysis but however for some factors, such as environmental, social, and political, are difficult to quantify in monetary term. The multi-criteria analysis is the alternative tool which caters the requirements of the issues above. The research, which is based on 102 class D and class E paved road sections in Kenya, is about to optimize road network investment under budget constraints by applying a multicriteria analysis (MCA) method and compare it with the conventional economic analysis. The MCA is developed from hierarchy structure which is considered as the analytical framework. The framework is based on selected criteria and weights which are assigned from Kenya road policy. The HDM-4 software is applied as decision-making tool to obtain the best investment alternatives and road work programs from both MCA and economic analysis. The road work programs will be the results from the analysis using both MCA and economic analysis within HDM-4 software to see the difference and compare the results between both programs. The results from MCA show 51 road sections need periodic work, which is overlay or resealing. Meanwhile, 51 others need rehabilitation or reconstruction. The five years road work program which based on economic analysis result shows that it costs almost Kenyan Shilling (KES) 130 billion to maintain the class D and E paved road in Kenya. Meanwhile, the MCA only requires KES 59.5 billion for 5 years program. These results show huge margin between two analyses and somehow MCA result provides more efficient work program compared to economic analysis.


INTRODUCTION
Road infrastructure has important role in the national development and is the main arteries for moving people and goods.Roads support the social and economic development process by open up access for the individuals and communities to the markets, employment, health, and education facilities.In terms of that importance, roads are positioned as one of the largest infrastructure assets in some countries which consequently need an adequate funding to finance, both constructions and maintenance.
Providing adequate budget is challenging for most of the countries, especially in developing and emerging countries, as the available resources are not sufficient to invest or even maintain the existing road network in good condition.The shortage of budgets in road investment results in high congestion, where the road length growth is less than vehicle number growth.Meanwhile, lack of budget for maintenance results decay of existing road networks.The portion of road funding in national budget itself varies five to ten percent of a government's recurrent expenses and ten to twenty percent of its development budget (Heggie and Vickers, 1998).This shortage of funds has been increasing travel time, vehicle operating costs (VOC), road conservation, pollution, and road accidents.
Planning and budgeting are two of the main keys in road network maintenance.Road budgeting is based on certain planning period or a multi-year program which consists of individual programs in an annual work program (AWP) package.These programs have to be realistic and practical to be delivered within the planned time frame but somehow due to constraints in financing, road budgeting has limitation and specified ceilings during planning period.Considering that constraint, prioritization in investing the budget is needed to achieve maximum impact and value-formoney on road maintenance expenditure (Odoki and Odongo, 2016).
Prioritizing the investment is complex as there are not just one factor that needs to be considered but also factors that are difficult to quantify or subjective.The method that can be implemented is multi-criteria analysis (MCA) which enables to consider not just economic criterion but also social, environmental, and political criteria.The considered criteria will be weighted and be applied to generate road networks program prioritization.

Road Network Management
Road network is the main vein in supporting social economic development, as well as environmental improvement.From this vital role, road network should be managed in appropriate and effective ways, in terms of its improvement and maintenance.The aims of road management are to enable the network to withstand the damage caused by wear and tear, to prevent sub-standard conditions, to ensure that traffic can use the network safely, efficiently and also concern about the environment (DFID, 1998).These processes are achieved through activities that have impacts on the road network.

Management Functions
Functions in road management need to be considered as the basis of management decisions will be made.The functions are focused from the planning through the operations with different scope of management aims, network coverage, time horizon, management staff concerned, and data level.Robinson (2008)  The key aspect of programming is to prioritize road works to find the best value for money in case of a constrained budget.Typically, programming develops budget for annual work programs for the road network within medium term time horizon, on year or multi-year programs.

c) Preparation
Preparation is the phase to packaging schemes and projects, which are selected from programming phase, into implementation.In this phase, budget usually has been approved as design specifications and costs are feasible to be implemented in the final schemes or projects.

d) Operation Management
Operation management covers on-going works activities which focus on individual sections or sub-sections of road.This phase is the detailed and implementation process of the works so the decisions are made in immediate time, typically daily or weekly basis, to monitor and to control the project activities.

Management Cycle
Road management is a process which integrates the cycles of activity involved in the management functions of planning, programming, preparation, and operation management (DFID, 1998).In performing those management functions, a logical sequence of steps need to be defined then the detailed content in each step depends on which functions will be addressed.Road management cycle as shown in Figure 1 will be applied to within road management as guidance in decision making with accurate information as the core of the cycle and typically completed once annually or in one budgeting period.
Figure 1.The road management cycle (Robinson, 2008) This management cycle can be utilized to support decision-making in every management function and is undertaken once for each period of the function (Adiguna, 2017).The description of each step depends on which function of management cycle is addressed.Every step in this cycle is related to the road network data as the core of the cycle.

Decision Support System
Decision support system is applied within the management functions to assist the tasks within.It divides into three main hierarchies, which are strategic, tactical, and operational.Decision support system basically assists in process, analyze data, and produce outputs from the analysis.This approach is the basis of the HDM-4 which will be applied in this research.
In programming level, tactical decision-making is applied to select the most optimal investment alternatives for the road networks.In order to optimize the decision-making, there are several things to be considered (Robinson, 2008): a) Consistent treatment selection methods across hierarchies in the network which are recognized for most defects.b) Condition projection methods to predict the performance of proposed treatments.c) Prioritization methods to get the optimum treatments for the road network under a specified amount of budget.
In making a decision, there are situations when budget is becoming the main constraint which is not sufficient to fund all of the road projects.In this situation, it is important to undertake a method to support the decision making in maximizing investments benefits from the limited budget.
Optimization is the process solution by finding the best way of using available resources while not violating any of other constraints.

Application of Highway Development and
Management System (HDM-4)

HDM-4 in Road Management
The Highway Development and Management System (HDM-4), which was developed by the World Bank and other international organizations, is one of decision support tool to evaluate investment options in road infrastructure.It is a road investment appraisal tool that contains individual components of total transport cost within analysis period.This accomplished by combining those components, such as environment, construction standards, geometric standards, and traffic characteristics.

Analytical Framework
The analytical framework within HDM-4 is based on the concept of pavement life cycle analysis and applied to predict road deterioration, road works effects, road user effects, and socioeconomic and environmental effects within typically period of 15 to 40 years depending on the pavement type (Odoki and Kerali, 2006).HDM-4 will predict the life cycle pavement performance under specified scenarios of road maintenance and/or improvement and the resulting agency and user costs.
HDM-4 is designed to make comparison of cost estimates and economic analysis of different investment scenarios within the homogeneous road section.The scenario consists of two or more different road maintenance and/or improvement works for each road section with one option designated as a base case (minimal routine maintenance).

Economic Analysis
Economic analysis is used to give a comparison in economic viability of different alternatives, to provide criteria for economic decision making, and also to investigate technical standards and strategies to be applied in the investment decision.In analyzing economic factor, there are indicators need to be considered to help make decision over the alternative of the projects depend on the characteristics which can be seen in Table 1.

Multi-Criteria Analysis
Multi criteria analysis is a simple analysis using criteria as attributes assessment can be either qualitative or quantitative assessment.Broadly speaking, the MCA activity consists of several main steps namely: setting goals, criteria, weighting criteria and scoring alternative decisions relating to the criteria (Aprilischa, et al., 2015).Cafiso, et al. (2002) defined MCA as a method which provides a systematic framework for breaking a problem into its constituent part in order to understand the problem and consequently arrive at a decision.MCA can assist the decision maker in road management to pick the best alternative as some of the criteria are not quantified in monetary terms then it combines both quantitative and qualitative criteria into single analytical framework.
The types of MCA methodologies have difference in terms of the preferences on the various criteria are specified, in ranking the alternatives, and also in presenting the results of the analysis.From all of the methodologies, the Analytical Hierarchy Process was selected because it produces a multi-criteria ranking index for every alternative as a measure of utility which is more compatible to be implemented in road maintenance management (Caviso, et al., 2002).
The outputs of MCA process will be multi-criteria ranking for each alternative in each section of the road network for each year of the analysis period.The alternative with highest value multi-criteria ranking number will be chosen for each section.If budget is constrained, prioritization or optimization needs to be considered by applying further analysis.
Multi-criteria ranking number will be used in optimization of the program.The value is calculated within HDM-4 as the ratio of utility index to the cost of implementing the alternative.The utility index can be considered analogous to the NPV used in economic analysis and can be used as an indicator for ranking and selecting projects.

RESEARCH METHOD
This research starts with identifying aim and objectives then followed by reviewing some literature to understand the concept of road management.The next step is defining data requirement then collect and collate those data.Analyzing data will be based on two methods, both are multi-criteria and economic analysis as comparison, using HDM-4 software.
The economic analysis is used to define the program analysis and prepare the multi-year work programs under budget constraints.In this term, the prioritization is based on NPV/cost ratio as ranking index with analysis procedure is similar to the project analysis as defined before.
The life-cycle analysis was selected as the method to run the program with period of five years.The two scenarios were undertaken, the first one is working program with unconstraint budget scenario and the second one is constraint budget scenario.Then the analysis results will be used as basis data to make road works program under budget constraints.

Data Requirements
Data requirements for MCA will be split into two stages.The first one is the data for HDM-4 inputs and the second one is data to determine relative weight for the MCA.
The data for HDM-4 is based on Information Quality Levels (IQL) related with the level of detail required for each application level (Robinson, 2008).Based on IQL levels, this research conducting a road works program with medium term time phase is defined at IQL-III and IQL-IV for the data needs.The detailed for HDM-4 input data which are required are road network characteristics, vehicle fleets, traffic, and road works.
Another required data is weightings in MCA.The roads policy has main aim for Developed Connectivity.The main aim is supported by these following objectives, accessibility, safety, economy, quality, congestion, and environmental impact and will be achieved by determining work alternatives for each object based on criteria on HDM-4.

Data Collection and Collation
The data for HDM-4 is obtained from consultancy project data which is based on Kenya road work project.

Road Network Characteristics
The details of the network are described as follows. a

Vehicle Fleets
The relevant motorized vehicles were defined with eleven vehicles as the representatives in Kenya HDM-4 workspace.Equivalent Standard Axle Load Factors also has been defined for each fleet based on axle load data from weighbridges in Kenya.

Traffic Data
Traffic data were defined from the consultancy data.The traffic growth rate data was inferred based on historical records of vehicle-km for traffic within the corridor in Kenya.Table 4 below shows the traffic growth for the analysis.

Road Works Data
In this research, maintenance work standard was defined for each road section.The maintenance standard consists of three alternatives as shown in Table 5.This maintenance work standard intervention was developed based on the time interval, the traffic level, and pavement condition itself.3).The following paragraph defines the methodology in developing the structure.

a) Define the main goal of the analysis
The goal in this research defines as Developed

Connectivity. b) Define the objectives
There are six objectives that have been defined.Those objectives are accessibility, safety, economy, quality, congestion and environment impact.c) Define second level objectives These objectives define from MCA criteria that supported in HDM-4 analysis.The criteria selected are related to the previous objectives defined before.Those criteria selected are maximised social benefits, reduce accidents, minimize RUC, maximize NPV, good riding quality, reduce congestion, and reduce air pollution.d) Assigned hypothetical weights for each criterion The weightings are important to determine the importance of each criterion and based on Kenya road policy which reflects the preferences of decision makers.For this case, RUC is selected as base criteria and the rest criteria are weighted based on the base criterion.The relative weight is shown in Table 6 and the relative weightings for each criterion are shown in Table 7. Major satisfaction e) Assigned alternatives for each criterion The alternatives are based on road works which have been defined before.Three alternatives, routine, periodic, and rehabilitation works, are selected for each road section.Net benefits to society (NPV) The net present value is calculated in HDM-4 for each investment alternative.The performance index to show the achievement of an objective to maximise benefits to society is the NPV.c)

Safety criterion
The number of road accidents by severity is calculated internally within HDM-4, for each investment alternative.d) Riding quality This criterion also calculated within HDM-4.The comfort attribute is the ride number (RN) and is based on the IRI for each investment alternative.This criterion is to show the road functional service with respect to comfort.e) Road congestion This criterion is defined by the volume-capacity ratio, VCR, which is calculated internally within HDM-4 for each investment alternative.The performance index is aimed to show the achievement of the objective to reduce road congestion.f) Environmental Criteria Quantities of different types of pollutants (vehicle emissions) are calculated internally within HDM-4 for each investment alternative.The environmental impact in terms of air quality index for each investment alternative is based on the concentration of hydrocarbon, carbon monoxide, nitrous oxide, carbon dioxide, Sulphur dioxide, particulates, and lead.Those pollutants have been defined in the HDM-4.g)

Social
The attributes required to calculate the performance index to show the achievement of an objective to maximise social benefits to society are not calculated within HDM-4.Therefore, it has to define the performance index for each investment alternative, based on their own judgment, by choosing from the options given in Table 7. h) Outputs The MCA procedure described above will produce multi-criteria ranking numbers for each alternative of each road section included in the study.The alternative with the highest value is selected for each section.

Analytical Framework
Analytical framework has been developed as a hierarchy structure as mentioned before.It consists of four levels of structure from the main goal at the top level, followed by the objectives then second level objectives where the weighting is assigned for each criterion.The last step is assigning road works alternatives for each criterion.The result of the developed analytical framework was applied in the analysis using HDM-4 software.Meanwhile, the hierarchy structure of the analytical framework is shown in Figure 4.

Multi-Criteria Analysis
The MCA is proceeded to evaluate the best alternative of three alternatives for each road section in accordance with the relative weightings in each of seven criteria, which are maximised social benefits, reduce accidents, minimize RUC, maximize NPV, good riding quality, reduce congestion, and reduce air pollution.The analysis starts in 2016 with 15 years period time and discount rate at 10%.The sample result of MCA is shown in Table 8.For every section, the alternative with the highest multi-criteria ranking number is chosen as the most suitable work.
The result from MCA analysis defined as the preferred road works alternatively for each section which has the highest multi-criteria ranking number.This preferred alternative can be source for the unconstrained works program.Multi-criteria value for each section and the most suitable alternative has the highest value compared to other alternatives in that section and highlighted to indicate.The result lists show different type of works and schedule for each road section.These are the effect of different budget scenarios and prioritization in the work programs.The prioritization here is based on NPV/cost ratio as the ranking index.The type of works for each section differs between periodic and rehabilitation works.
The results obtained from the analysis shows that for unconstraint budget scenario needs nearly KES 130 billion for five years period.For the constraint budget scenario was implemented with another two different schemes.The scenarios and the annual budget for each scenario are shown in Table 10.During the 5 years, road works program, the unconstrained budget scenario will maintain the road network from 2016 until 2019 and in 2020 will be minor road work such as routine maintenance.
Meanwhile, for the constraint budget scenarios, the major road works will only be held from 2016 until 2018 and the rest will only be maintained by minor road work.
The budget required for the MCA results show different value.In this term, the budget requirement for MCA is lower than the unconstrained economic analysis.The MCA unconstrained works program requires KES 59.5 billion for five years work analysis.The annual budget needed for MCA works program is shown in Table 11.The different budget requirements may happen since the two analyses have different approach to assign the requirements.The Economic analysis uses pavement condition to undertake the program and the MCA considers preferences of relative weights that have been assigned before which may satisfy the high weighted criteria.
In The economic analysis is also applied to define the program analysis and prepare the multi-year work programs under budget constraints.The results obtained from the analysis shows that unconstraint budget scenario needs nearly KES 130 billion for five years period.For the constraint budget scenario was implemented with another two different schemes, each 80%, and 60%.If it is compared with MCA road works program, the economic analysis program has higher allocation with huge margin between them.The difference in allocation effects in the program prioritization and type of road works for each road section.This condition may happen as MCA road works program accommodates more criteria to be considered rather than just economic criterion.For Kenya road works program, MCA is more efficient and giving more analysis options for the road sector stakeholders there.

Recommendations
Multi-criteria analysis has more efficient road works program if it is compared to the conventional economic analysis but further studies are required, such as: This analysis is applicable in Indonesia as it enables wider factors to be considered.Just like in many countries, Indonesia is also facing an inadequacy of road funding.Directorate General of Highway in Indonesia could implement MCA to create efficient national road works program in Indonesia.The analysis in HDM-4 would be based on the road networks in small working area then it is compiled into provincial program and national program.
Somehow it is challenging to implement the MCA in Indonesia national road works program as Indonesia has long road networks with different characteristics in each region and difficulties in data collection.Therefore, it needs more efforts for the Directorate General of Highway to implement this method as comprehensive and detailed data are important to get an accurate and efficient road works program.

Figure
Figure 3 Hierarchy structure development

Figure 5 .
Figure 5. Annual budget comparison a) Combining high and low-volume road to see MCA affects the real road works program as this research only analyses low-volume roads (Class D and E) in Kenya, b) Comparing final road condition resulting from MCA and economic road works program after five years.The MCA road works program has significant lower allocation so it is better to check the road condition after the multi-year programmed has finished, c) Assessing more comprehensive data with the road sector stakeholders to determine the weights of each criterion in MCA as stakeholders have significant roles in the decision making.

Table 8 .
Sample of MCA results

Table 9 .
Number of road sections in MCA results

Table 11 .
MCA annual costs class D and class E paved road in Kenya in five years analysis.Meanwhile, the MCA only requires KES 59.5 billion for 5 years program.These results show huge margin between two analyses.This resulted in different schedule and different type of road works for each road section.