Rural Road Network Decision Model for Hilly Regions of Nepal

Abstract

Transportation cost is one of the major costs for public and private sectors in rural areas of developing countries (in order to deliver goods and services). This is due to the absence of links or to their poor physical conditions (such as road surface in a road network). To reduce the operation costs in existing links better road surfaces are needed. However, only some links can be improved or constructed to the best surface level at the same time, due to resource constraints. Hence, a careful decision about which links should be improved or constructed to achieve the minimum transportation cost is needed. This paper presents a decision support model for a rural road network that provides portfolio of suggested links for road network improvements or constructions and offers solutions for different budget levels minimizing the transportation cost in the rural road network. Mixed integer programming is used to get an optimal solution.

Roads Development Optimization for All-Season Service Accessibility Improvement in Rural Nepal using a Novel Cost-Time Model and Evolutionary Algorithm

“…The accuracy with which models reflect seasonal access changes depends on the model and metrics employed. Transport geographers (Jafino et al, 2020;Rodrigue et al, 2016;Jenelius, 2009) and civil engineers (Shrestha, 2018) have developed concepts of criticality and redundancy to explain the impact of seasonal disruptions on road network performance within a network analysis. Because these network-scale analyses approach seasonal impacts in terms of binary open-closed changes in roads' status they obscure quantitative changes in network and road-scale performance from reduced speeds over damaged but still-functioning roads.…”

“…compared to the others. This section discusses the optimization approaches which exist to identify these superior alternatives and how they have been employed in works concerning rural roads Optimizing across rural networks presents a unique and understudied challenge since traditional transportation optimization models are designed for urban, relatively developed settings and hence are likely to underperform in providing solutions for rural, sparsely populated settings (Shrestha, 2018). However, the optimization of rural roads development has not been entirely overlooked and has seen recent advances, building upon earlier foundational work in this context by Kumar and Tillotson (1991), Kumar and Kumar (1999), and Kumar and Jain (2000) -noteworthy is that these studies are all focused in South Asia.…”

“…The model then approximates solutions which provide optimal trade-offs in these objectives, determined for different maximum budget values. Similarly, Shrestha (2018) presented a multi-objective decision support model for rural road networks that suggests road network links for improvement or new construction. Using a similar approach to Shrestha et al (2014) and with the objectives minimizing four population-and distance-based criteria, optimal trade-off solutions were approximated at different budget levels.…”

“…Heuristics are considered because they are easily compatible with rural roads planning (Ma et al, 2014;Scaparra and Church, 2005;Kumar and Kumar, 1999) and can obtain numerous well-performing solutions within reasonable computation times when the complexity of the problem is too overwhelming for traditional methods. An integer-linear programming (ILP) approach is the second method that we implement, because it provides a simpler way to solve optimization problems compared to our heuristic approach and has been used in rural roads planning before (Murawski and Church, 2009;Shrestha et al, 2014;Shrestha, 2018). An ILP approach is, however, more sensitive to computational complexity and the size of problems which it can solve, and returns fewer solutions when compared to a heuristic approach -resulting in less data and less insightful solution-analyses.…”

“…This weighted-sum method provides a simple way to solve MO optimization problems, since the mathematical formulations are relatively simple to provide as input compared to the sophisticated code and multiple parameters and sensitivity tests that are required for heuristics. Due to its accessibility and historical prominence the weighted-sum approach remains a popular choice to solve MO optimization problems from various recent applications, such as building design (Machairas et al, 2014), carpool matching (Xia et al, 2019), and land-use allocation (Yao et al, 2018), and has also been used for rural roads planning research (Shrestha et al, 2014;Shrestha, 2018). A popular ILP alternative to the weighted-sum approach is the ǫ-constraint method (Mavrotas, 2009;Mavrotas and Florios, 2013), which has been shown to be capable of obtaining superior solution spread along the Pareto front approximation in certain problem instances Amin and Zhang (2013).…”