Prof. P.K. Sikdar

INTRODUCTION

The connotation of rural roads for a developing country like India is different from that in the Western countries, in the sense that, it is not the road which passes through rural and agricultural areas. It is because of the fact that a road passing through rural area can be a National Highway, State Highway and MDR (Swaminathan et al, 1980). As per the definition and classification of road system adopted in Road Development Plan of India (1981-2000), rural roads are the tertiary road system which comprises of Other District Roads (ODR) and Village Roads (M.O.S.T., 1984). These roads serve as feeder roads to the main network and link villages with the nearest market centers, or any other village. Traffic volume on these roads is low, comprising mainly of slow moving vehicles like cycles and animal drawn vehicles and pedestrians who are generally cultivators of abutting land. There are other lower level roads also, such as paths and tracks, which serve the important functions in rural areas and are not covered under the above definition and classification of roads. These are farm roads and intra-village roads. The farm roads traverse the agricultural fields and are mainly unsurfaced linking agricultural farms with the main village. Intra-village roads connect the clusters of small size settlements in a village, known with different names in different states (Srinivasan, 1985).

All these types of road constitute the whole range of rural roads for all practical purposes in India. However, from network planning point of view rural road is an all weather road that connects a village with any other village, to the market centers, or to the primary and secondary road systems. This constitutes the Other District Roads and Village Roads as per the definition of Road Development Plan (1981-2001) of the country. The aim here is to develop the network planning methodology for this category of village roads using GIS database. A case study of Taluka level rural road network planning, which is an intermediate level between a village and a district in India is presented here. This signifies the typical approach of rural road planning that should be adopted for a developing country. But, before that the requirements of rural road planning, different approaches of rural road planning in India, and the rural travel characteristics are discussed.

TECHNIQUES OF RURAL ROAD PLANNING

The ultimate objective of solving this system is to provide road connectivity at appropriate level of serviceability to serve as much of population as possible, with involved road length kept to a minimum. The rural roads for the inter settlement interaction are required to fulfill their missing socio-economic functions. The road network, therefore, should match the movement pattern in the region. The rural road

planning practice in the country is guided by the 20 year Road Development Plans. The guidelines given in these plans are nothing but rules of thumb for providing connectivity to village with the existing road network based on their population. Thus, the present state of the art in respect of rural road planning and implementation can most appropriately be described as traditional in character. There is not much change in age-old practice and procedures followed in this regard. There has been some research carried out in recent past towards planning the rural roads in a scientific manner, considering the socio-economic need of the rural population. However, they have not yet been implemented to a wider area across the country. The following sections, therefore, reviews the existing practices of rural road planning, identifies their weakness and then suggests the research approach to tackle the rural road problem in the country.

Road Development Plans

Since 1943 three successive 20 Years Road Development Plans were formulated and two of them were executed. The third 20 Year Road Development Plan (1981 – 2001) is in operation and the provision of roads in rural areas is being made as per the guidelines set forth in it. As per the objectives of this plan all the villages with population over 1500 should be connected by Major District Roads and the villages with population 1000 to 1500 by Other District Roads. Thee should be a road within a distance of 3.0 km in plains and 5.0 km in hilly terrain connecting all villages or groups of villages with population less than 500. the target road length in this plan was calculated to achieve a road density of 82 km road length/100 Sq. Km of area in the country. There are many shortcomings in adopting these guidelines as the basis for planning rural roads. Firstly, the road mileage targets fixed in the plan are based on certain empirical formulae. Instead of formulating the road development plans at local or regional level, pre-calculated road mileage is assigned by the plan, and states and districts are required to reconcile their needs with the overall target thus assigned, without considering the actual accessibility needs of a region (Mahendru et al., 1983.) Secondly, there are no guidelines to decide inter settlement connectivity. The master plans prepared by states for road development based on these guidelines result in direct connectivity of villages of qualifying population range with the nearest existing road. Network thus developed may not be the most efficient catering to the functional requirements of settlements. This approach has thus resulted in development of sub-optimal network and poor utilization of scarcely available resources in the country.

Network Planning Approach

The research approach to develop the rural road network in the country can be broadly classified into two main categories. The first category is mainly based on the concept of minimal spanning tree and the other is based on inter-settlement interaction approach. These two approaches are critically reviewed here.

Swaminathan, et al. (1981) used the concept of minimum spanning tree for connecting the villages to existing nearby roads or to the market center. Various link options for connectivity were analysed by considering the flow circuit. Market centers and existing roads were considered as high intensity concentrated electric charges,

which attracted smaller charges situated at unconnected villages. The weight of each link option was taken as the ratio of electric charge situated at the already connected village or to market center or the nearly road and distance between them. The option giving the highest weightage was chosen for providing connectivity using minimum spanning tree concept. In this method it was claimed that the network generated is the optimal but no analysis for optimality was done. The charges assigned to market centers and main roads were chosen arbitrarily. The logic of the assigning charges to the road and then finding it for other villages is a misconception. For an unbiased linkage scheme, the road should receive charges from the village and not vice-versa.

Kumar and Tilloston (1985) proposed the rural road network planning methodology which minimizes the road construction and travel cost. The villages were considered as “unconnected nodes” which were to be connected to “root nodes”, situated either on market centers or on the existing main roads interconnecting the market centers. The minimum construction cost network was generated first by using minimum spanning tree concept. So the unconnected villages were connected to the market center or the main road and proceeding towards interior by connecting the nearest unconnected village with the already connected ones. Alternative networks were generated from a set of predetermined road links using different link options. The optimum network was obtained from minimum construction cost. In this method it was assumed that all the functions of a village are satisfied if it is connected to a market center, which is not true. The set of link options, from which the alternative networks were generated, were a predetermined few. Whereas, if there are ‘n’ nodes in a network then the maximum possible number of link options are n(n-1)/2. Not all these link options were considered in generating the alternative network. The method suggested did not consider an integrated area development approach, and therefore, its overall functional utility was questionable.

Mahendru et al. (1982, 1983, 1985, 1986, 1988) used the concept of settlement interaction, link efficiency, route efficiency and network efficiency to generate, analyse and evaluate alternative rural road linkage patterns. Integrated area development approach was considered to arrive at a road network, which serves the area in a balanced way. Gravity hypothesis was used to qualify inter settlement interaction using level of socio-economic development, population and spatial in terms of centrality scores and the interaction between two settlements was considered proportional to the difference in their centrality scores. Alternative networks were generated using various criteria like maximum link efficiency, minimum total link length, minimum total operating cost and fully developed network. These alternative networks were tested for their total cost, which consisted of construction and operating costs, to arrive at the optimal network. This theory is based on the most scientific and rational way to tackle the rural road planning in the country, until date. In a true sense it attempts at integrating the rural road network planning process with the over all development of the area. However, a close perusal at the modeling philosophy adopted raises some doubts in the approach. Firstly, the Gravitation hypothesis, used to model the hidden pattern of inter-settlement interaction, is itself questionable. The gravitational model used to find the demographic force of inter-settlement interaction is as given in the next page.

Fij = Pi X Pj (I CSi – CSj I)/d ij

Fij = Force of interaction between settlement i and j

Pi and Pj = Population of settlement ‘i’ and ‘j’ respectively

dij = spatial distance between i and j respectively, based on level and number of socio-economic facilities available at settlements.

As per the above equation if Csi and CSj are equal for any two settlements, then the force of interaction becomes zero which is only possible, when the number and type of functionalities at both the settlements is identical. Since Csi and CSj are the aggregated score arrived from various functionalities present at the settlement, their values can be equal even if the facilities are entirely different. Therefore, under these circumstances, the force of interaction found from the above formulae may be misleading. Primary and secondary road systems of NH, SH and MDR were not taken in the framework for planning the village roads. By not considering the primary and secondary road systems during the planning process the possibility of connecting a village through the existing road network was completely ignored, which is another serious limitation.

Kumar (1997) suggested the facility-based approach to rural road planning. One of the important contributions of the study is its findings about the rural travel characteristic, which were derived from an extensive survey data obtained from rural areas. It was observed from the survey that 92.79 present trips terminate either at market center or education center. Therefore, it was suggested that, a network, which provided connectivity to market center and educational institutions in an correlated with their accessibility from different road types. Education level was taken as the proportion of population studying or studied at a particular level whereas the accessibility measure was taken as the distance of education institutions from the village. The existing correlation between accessibility and education level was considered as the guiding tool to arrive at the maximum permissible distance of the village from tan educational institute in planning the rural road network. The network design approach was to connect a village to the nearest market center so that maximum number of educational institutions fall along the designed linkage pattern or at least they should satisfy the maximum distance criteria for the educational institutions. Although, in this study the trip pattern in a rural area was captured precisely it was not used efficiently in the network planning. The network planning approach was to extrapolate the existing accessibility pattern of villages to the unconnected villages in developing the linkage pattern. The total accessibility requirement of village was not considered in arriving at the optimal network.

RURAL TRAVEL CHARACTERISTICS

The mobility and transport infrastructure details for some of the developed countries in the world are given in Table-1 (Simon, 1997). Here the overall travel mobility, is estimated as passenger miles per year by automobile, rail and domestic air travel. It is evident from the table the extremely low mobility is the result of very poor transport infrastructure in India in comparison to developed countries of the world. These mobility parameters declines further if only the rural areas of the country are

considered. The wheeled vehicles in the rural areas in the country are mainly cycles and animal drawn carts. Among these, the proportion of motor vehicles is still very low.

Based on the findings of rural travel survey (Kumar et al 1997) it was observed that nearly 50% of the population in rural area is static i.e. they do not make any trip. The highest number of trips in rural areas was for education (42%). The percentage of trips, classified according to their purpose is as shown in Fig. 1. it was observed that nearly 86% of trips, to other than educational trips terminated at the nearest market center. The trips either at the market center or at the schools constituted 93% of total trips. This suggests that school and market are the main activity centers in the rural areas in the country.

METHODOLOGY FOR RURAL ROAD PLANNING

It is evident from the above discussion that the rural travel pattern in India is entirely different from those in developed countries. Therefore, the transport planning techniques and innovations suitable for developed economies of the world are not appropriate for planning the rural road network in this country. In other words, the conventional transport planning process, which is based on the principle of subdividing the study area into smaller zones and sub zones and studying the interzonal trip frequencies through OD matrix to decide linkage needs between various origins and destinations, are not of much use. Even planning at the district level, there will be around 1500 to 3000 settlements and it will require enormous amount of resource to collect trip frequency data for all the settlements to form the O-D matrix. On the other hand, the observed trip frequencies will be very low which the relative priorities of the linkages (Mahendru et al. 1985). As it has been pointed out earlier, the rural road planning should be based on integrated area development approach considering the spatial distribution of various functions in a region. In addition, the rural roads should be planned in such a manner that they serve the functional requirements of each settlement in an efficient manner. In other words, the linkage pattern developed should provide accessibility to various functions available in a rural area, to all the settlement in an efficient manner. Therefore, in a true sense the village road planning should be based on the accessibility requirements of the settlement for their missing functions. In the present study a methodology for rural road planning has been evolved which is primarily based on the accessibility concept.

4.1 Accessibility Approach for Rural Road Planning

The basic philosophy behind accessibility based approach for rural road planning is that most of the unconnected settlements in rural area dependent on some or many other connected settlements in their neighborhood for their missing functions.Due to these missing functions, the settlements interact with each other in the form of movement of men and materials or in the form of travel for different purposes. A most suitable road link, which connects an unconnected settlement with either a connected settlement or the existing road network, is the one which provides maximum accessibility to the unconnected village for its missing functions.

Therefore, in the present approach the rural roads are planned in the context of an already existing network of primary and secondary road systems.

Design of accessibility indicator

The main objective of rural road planning here is to provide connectivity of the unconnected settlement to the existing road network through at least one link. This link may be either connecting to the connected settlement or to any other node on the existing road network. Therefore, for rural road planning purpose the entire system of nodes can be divided into two categories such as, unconnected nodes (settlements) and other nodes on the existing road network. The second category is comprised of connected settlement nodes and the nodes on the links in the network. In order to analyse the connectivity of unconnected settlement with the nearby road links additional nodes on the links of the network are created at some appropriate interval. This is shown in Fig.2. if the total number of nodes on the network is, ‘m’ then there will be these many link options for each unconnected settlement. The link option, which provides maximum accessibility to the unconnected settlement, is chosen in developing the linkage pattern. An accessibility index is designed to evaluate these link options, which is explained in the following.

Let us assume (refer Fig.2) that

‘i; denotes and unconnected settlement

‘j’ denotes nodes on the existing network

‘l’ denotes the link option between unconnected settlement ‘i’ and node ‘j’

1, 2, …., k are the various functions, a set or subset of which is present in various settlement.

dkl i is the distance of facility type ‘k’ for unconnected settlement i by link option l.

dk j is the distance to nearest facility type ‘k’ from any node ‘j’ on the existing network to other already connected nodes.

then

dkl i = dkj+dlij

where dlij is the length of the link option l between settlement ‘i’ and node ‘j’ on the existing network.

Here, it must be mentioned that if the function ‘k’ is present at settlement ‘i’ then

dkl I = 0, and if it is available at j, then dkl ij

if the total number of trips to facility type ‘k’ from the unconnected settlement ‘i’ is Tk i then total person-km of travel for settlement ‘i’ to all possible other settlements through link option ‘l’ will be

PK li =?k Ti k. di kl

If the mass (i.e. population of unconnected settlement ‘i’ is Pi then accessibility index of settlement ‘i’ through link option ‘l’ can be defined as:

Ai l = ?k dikl.Tik/ Pi

Here Ail is the accessibility index of settlement ‘i’ through link option ‘l’, which is the composite measure of accessibility of the link option connecting any unconnected settlement ‘i’ to any other node on the network. Higher the value of this index lower is the accessibility and vice-versa. This accessibility measure is based on the assumption that it the unconnected village ‘i’ is devoid of facility ‘k’ then by connecting it to node ’j’ will satisfy its requirement with the nearest location of facility ‘k’ at node ‘j’ or elsewhere in another settlement. This indicator represents the average person lead, for an unconnected settlement, in fulfilling its missing function using a particular link option. It is obvious that the link option, which ahs minimum value of this accessibility index (i.e. maximum accessibility), is selected in developing the linkage pattern. To find the accessibility index according to equation (4) the concept of potential trips is used for estimating Ti k. Potential trips are derived from the trip rate fro each function type. There are many intricacies involved in estimating the term dkl i. These issues are dealt in the following sections.

Potential trips and trip rate

The number of trips Tkj in equation (4) is not the present total number of trips performed but it is the total expected number of trips as per the developmental policies of the area. In this sense, it is the total number of potential trips corresponding to function ‘k’. in absence of any model, at present, to estimate the total trips of type ‘k’ for rural areas, the concept of trip rates can be used to find Tki.

Tik = Pi. Trk

Where Pi = the mass (i.e. population, number of households etc.) for settlement ‘i’ Trk = the trip rate for the purpose of function ‘k’.

In rural areas appropriate models should be developed which correlates the trip rates for each function ‘k’ with the development level of the settlement. The model can be a function of the socio-economic factors. However, Trk is quite easy to arrive at for a few of the functions. For example, if we want that all the rural population should be educated up to 10th standard by the end of a chosen planning horizon then the trip rates to middle school and high school will be nothing but the proportion of population in their corresponding age groups. But, the trip rates for marketing, health purpose, etc will be governed by the socio economic level of development and the future developmental policies of the area.

Alignment of link options

One of the main objectives in rural road planning is to minimize the construction cost of new link which depends on its alignment. In other words the approximate construction cost of each new link option should be known before hand, for considering it in planning process. Two things are required to achieve this. Firstly, the alignment of new link should be chosen in such a manner that it encounter

minimum number of natural barriers like rivers, hillocks, ponds, lakes etc. in addition, the land acquisition cost and the amount of cutting and filling should be minimum. There can be many other such aspects, which should be considered in the alignment of any new road link. Secondly, the costly link options should be discarded in developing the linkage pattern. However, it is a difficult task to suggest the least cost alignment manually, which depends on so may topographic factors. This becomes further difficult for a set of ‘n’ nodes, where the total number of link options are n(n-1)/2. manually this becomes impossible even if all the necessary topographic details about the area are made available. In the present approach a methodology is devised based on the concept of grided network analysis, which can be accomplished through spatial analysis functionalities available in GIS. This analysis consists of overlaying the entire study area with small size grids. Since GIS handles separate layer corresponding to each topographic feature of the area, so each link of the grid can be queried spatially with respect to each topographic feature. After this query, the element of the grid gets loaded with the topographic information. This type of grided network along with a few topographic features is shown in Fig3. Here it is important to note that the superimposed layers of topographic feature can be either an area layer like lakes and hillocks or it can be a line layer of rivers, canals or even the existing roads etc. Now each element of the grid is queried spatially, using GIS functionalities, with respect to each area and line layer. After this query the each grid element gets the information of the number of cross drainage structures of different types. Similarly after querying them with respect to all other features some of the links can be sieved out as infeasible links. These link can be selected and disabled at the time of the alignment of new link options to ensure that it does not pass through these links. There can be many other such scrutiny based on land type, soil type, terrain type etc, which can used as the base layer to query the elements of the grid. The query with respect to these layers provides the additional information about the construction cost, land acquisition cost etc. for each element of the grid. After making these queries, each element of the grid gets the information about various topographical characteristics of the area and then the problem of road alignment becomes quite simple. The shortest path principle, taking the construction cost as the link attribute of grid network, can be used to arrive at the least cost alignment between any two points in the area. The accuracy of the alignment depends on the size of the grid elements. Smaller the grid size more accurate is the alignment. However, small grid size results in huge network and therefore the required computational power is also increased.

NETWORK GENERATION PROCEDURE

It is now possible to develop the rural road network using various criteria in selecting the best link option, which provide connectivity to the unconnected settlements. This can be the accessibility criteria as discussed in previous section, or some other criteria based on the developmental policies and objectives of the region. Some of these networks are discussed in following sections.

Maximum accessibility network

The accessibility index developed in previous section can be used for developing the network. This network is the most efficient in serving the missing functional

requirement of unconnected settlements. The step procedure to develop the network is given below:

• Classify all the settlements in the region into two categories i.e. connected and unconnected. The connected settlements are the set of nodes on the existing road network or having connectivity.
• Considering the existing roads as the network the distances of missing functions, if any, (eg markets, schools etc.) for each connected settlement is found. This is determined by analyzing the shortest path distance between all the connected settlements and the functions present in each connected settlement.
• Find the potential trips of each category for every unconnected settlement.
• Find the link options from every unconnected settlement, using link alignment procedure.
• Using equation (4) the accessibility index of each link options from each unconnected settlement is found. The link option which has the maximum accessibility index, should be selected for each unconnected settlement.
• Among these link options of various unconnected settlements choose the one which offers highest accessibility (i.e. the lowest accessibility index value). Therefore, the first unconnected settlement and its corresponding link option is identified for providing connectivity to the network.
• Find the distance of various facilities for the newly connected settlement.
• The accessibility index of the link options with this newly connected settlement to the remaining unconnected settlements, which are n-1, are calculated. The method repeats from step (iv) to obtain the connectivity of second unconnected settlement.
• Repeat step (iv to (vii) till all the unconnected settlements get connected.

Minimum construction cost network

A minimum construction cost network is the one, which provides the connectivity to all the unconnected settlements so that the total construction cost of all the new links is minimum, it can be generated by selecting the link option of minimum length or overall construction cost. It may be mentioned that a shortest link need not be the least cost option. The maximum accessibility criteria is replaced with the minimum construction cost criteria in the network generation procedure discussed in the previous section. This network will be less efficient will be less efficient than the maximum accessibility network in serving the missing functions of the unconnected settlements.

Optimum network

The network generated using maximum accessibility criteria has the minimum operating cost for reaching to the missing functions, but its construction cost is higher in comparison to the minimum construction cost network is higher than the maximum accessibility network. Optimum network is one in between satisfying the both

criteria and it is difficult to generate an optimum network in its true sense. However, it is possible to generate different networks between above tow extremes and select the one, which meets the policy objectives of the two criteria. This is elaborated further in case of an application analysis presented subsequently.

Other networks

In a policy guided network generation, those link options which offer accessibility more than certain critical value can be added as the new links to the network. The network thus may have more than one link in providing the connectivity to an unconnected settlement. Many other networks can be generated based on the criteria of providing maximum accessibility to certain types of function rather than considering all functions together. These functions can be education, marketing, health etc. based on different policy objectives the criteria can be framed and the corresponding network is generated.

APPLICATION ANALYSIS

6.1 The Database

The accessibility based methodology developed for rural road planning is applied to the Paithan Taluka of Aurangabad district. The map of Paithan Taluka is shown in Fig. 4. the database required for rural road planning is prepared on GIS platform. The data items in the context of application analysis, are discussed in the following sections.

Connected and unconnected villages

The villages in Aurangabad district are kept in a separate point layer in the present GIS database. In order to test the methodology of rural road planning, all the rural roads (i.e. Village Roads and Other District Roads) were treated as missing in classifying the villages under connected and unconnected settlement categories. All the villages in the Taluka were queried for their distance from the existing primary road network which constitutes Major State Highway (MSH), State Highway (SH) and Major District Road (MDR). Those villages which were at a distance less than 0.5 km from these roads were treated as connected and rest al were taken as unconnected. The connected and unconnected villages and the existing MSH, SH and MDR network in the area is shown in Fig. 5.

Travel requirements of unconnected villages

As it was pointed out earlier, 92% of the inter-settlement interaction in rural areas is to satisfy their educational needs and various other socio-economic requirements, which are primarily fulfilled at the nearest market center (Kumar et al, 1997). Therefore, the important socio-economic functions considered in the present analysis are related to education, marketing and health purposes. The educational institution considered for settlement interaction are primary school, middle school and high school, in each village. A village is considered as the market center if at least one day in a week there is some marketing activities. The health facility is

considered to be available at a village if there is one private practicing doctor or a Primary Health Center located at the village. The location of these functions in the villages of Paithan Taluka is obtained from the village directory (Population Census, 1991) of population census. The number of villages having these facilities in Paithan Taluka are given in Table-2. it can be observed from the table that, except the primary schools all other facilities are sparsely located in the Taluka. The villages interact for these sparsely located facilities. In order to estimate the total trips for education purpose, the trip rate is taken as the proportion of population in their respective age group. Here, it is assumed that the entire population is to be educated up to certain standard in the future. In this sense, the proportion of population in the age group which goes to school of a type, is the trip rate to that school type. Therefore, it is the potential trip rate. However, the market trips are calculated using the actual trip rate to the market centers, which is obtained from the rural travel data (Kumar et al, 1997). The trip rate to health center is not available now so it is suitably assumed. All these trip rates are given in Table-3. Using these trip rates the total number of trips for different functions is estimated for each unconnected village.

Distance of facilities from each connected village

In order to find the accessibility index of link options distance of various facilities from each connected settlement are required. Those facilities, which are present in a settlement their distances are taken as, zero, the settlements interact with other ones in its neighborhood due to their missing functions. To find the nearest settlement where the missing function is located, first of all the shortest path between all the connected settlement, through the existing network, is estimated using the path module available in GIS. A module in C++ language is coded to find the distance of various facilities for each connected settlement. The distance of various facilities for some of the connected villages in Paithan Taluka, obtained from this procedure, is given in Table-4.

Grid and road network layer

The spatial details of the road network for Aurangabad district consists of MSH, SH, MDR, ODR, and village Roads (VR). A square grid layer of 1 km. size with intersecting diagonals is merged with this road layer. The attribute data of these two layer are also merged and the links of the merged layer of roads and grid have their separated identity due to different code assigned to each type. This layer is now queried with the drainage line layers, which identifies the number of cross drainage structures on each grid element. An equivalent cost, in terms of kilometers of road length, is added for each cross drainage type to obtain total cost for the link. So the total cost of each link is found in terms of length of new road to be constructed. Using this layer, a network is formed which can be used now for the alignment of link option.

Using this network the alignment of all the link options is determined. If the number of unconnected and connected settlements are n1 and n2 respectively then a shortest path matrix of n1 x (n1 + n2) size is generated for examining all the link options. The attributes of a few link option in Table-5. the alignment of some link options is shown in Fig.6. All connected villages are taken as the nodes on

the existing network of MSH, SH and MDR in developing the link options from the unconnected villages. However, it is also possible to consider the intermediate nodes on the existing road links in developing the link options but the high computational requirements prohibits to do so. Under this situation, the existing road link attributes are set to zero to align the link option for connecting a unconnected village to the nearest road.

Network Development Analysis

In the present method analysis the existing village Roads and other District Roads are treated as missing first, and then the linkage pattern is developed. The network generated after the analysis is compared with the existing rural road network. This identifies the redundant road links in the existing rural road network from rural travel point of view. This comparison also brings out the important road links in the existing system of rural road network, which can be upgraded to the all weather black top type.

It the base map is considered at 1:50,000 scale many inter-settlement roads will be possible to consider which are mostly foot tracks or earth roads. Since these road already exist, the land acquisition cost along these alignments will be minimum. Therefore, these roads are the favorable link options in the development of the linkage pattern. In order to take into account these link options, all categories of these existing roads should be merged with the grid layer. Then, the land acquisition cost factor should also be included in finding the total cost of the links of the merged layer. Therefore, the total cost attribute along the existing road links will be less in comparison to the grid links in the layer. Now, If the links are aligned taking this total cost as the network attribute then the link options will align more along the existing road links.

Since the present spatial data base of the area is at 1:250,000 scale, it does not have these inter-settlement roads of foot tracks and earth roads. So, only the existing Village Roads and other District Roads (ODR), present in the database, are merged with the grid layer. The land acquisition cost along these links are set to zero so that the link options align more towards these road links in providing the connectivity to the unconnected village. Then it becomes quite convenient to compare the existing system of rural road network with the newly developed road links.

Keeping these objectives in mind the network development analysis is carried out for the following two cases.

Case–1 : With land acquisition cost
(The land acquisition cost is taken as equal to the road construction cost.)

Case-2Without land acquisition cost

Three types of networks are generated for each case. These networks corresponds to, maximum accessibility criteria, minimum construction cost criteria and optimality criteria discussed earlier. Although, there is only one optimal network theoretically, the optimality concept is used loosely for excluding the link options of exorbitantly

higher construction cost. So the highest accessibility link option, which may have a very high construction cost, is ignored and the next best link option is explored. The procedure for generating the optimal network is given below:

• Find link option using the minimum accessibility index. If the has the minimum equivalent construction cost also, select this link option.
• Otherwise, find the construction cost for the minimum accessibility indeed link option. Let us assume that accessibility and cost are Amin and L respectively. Let us also assume that the construction cost and accessibility index corresponding to other link options from the unconnected village, sorted in ascending order of their construction cost are L1, L2, L3,……. and A1, A2, …… Al respectively.
• Now find the optimality factor for all the link options which is defined as given below: Optimality factor =(Ai-Amin)/(L*-Li)
• if the minimum optimality factor has a positive value which is less than certain critical value, then the corresponding link option should be selected as optimal. Otherwise, adopt the minimum accessibility index link option.

This critical value of optimality factor will depend on the trade off between accessibility and construction cost of the link option. In the present analysis this value has been taken as 0.5.

The maximum accessibility network and optimum network for case 1 and case 2 are shown in Fig. 7 to Fig.10. For each criteria of network generation the priortised list of links is obtained. The overall characteristics of three types of networks generated for Case 1 and Case 2 are given in Table-6. It can be observed from the table that optimum network generated falls in between maximum accessibility network and minimum construction cost network in terms of person – km. of travel. The total length of the maximum accessibility network is 20% higher than the optimum network in Case 1. This id due to the fact that maximum accessibility network has duplicative link lengths. In Fig.11 a portion of the network for both the criteria, for Case-1, is shown in a comparative manner. It can be observed that in the case of maximum accessibility network there are many direct links from unconnected villages. Whereas, for the optimum network the direct links are reduced. The designed link options in the optimum network connects through the existing road network.

Due to the extra land acquisition cost for the grid elements, in Case 2, the link options are aligned more along the existing Village Roads (VR) and Other District Roads (ODR). The maximum accessibility and optimum network for case 2 are shown in Fig. 9 and Fig 10, respectively. However, the maximum accessibility network has still more direct links than the optimum network in Case 2. Whereas most of the links of the optimum network connects to nearest existing network of MSH, SH, MDR or is aligned along the existing ODR and VR. It can be observed from Fig.10 that the optimum network for Case 2 is mostly aligned along the existing road network of ODR and VR. A comparison of the portion of network generated using Maximum Accessibility criteria in Case 1 and Optimality criteria in Case 2 is shown in Fig. 12. it can be observed that most of the direct link options of the maximum

accessibility criteria network in Case 1 is now aligned along the existing network of VR and ODR without much duplication of road length. Therefore, the optimum network generated for Case-2 can be now compared with the existing VR and ODR. The developed linkage pattern or unconnected villages which overlaps with the existing Network or VR and ODR is important from rural travel point of view, and therefore, should be upgraded to all weather black top type. The comparison of the developed linkage pattern with the existing Village Road and Other District Road also identifies the redundant links in the existing network, if there is only one road connection to be given to each unconnected village. The portion of the existing network of VR and ODR, which does not overlap with the developed linkage pattern, is the redundant portion of the network as shown in Fig.10.

ACKNOWLEDGEMENTS

The material contained in this lecture note is entirely taken from the Ph.D. Thesis submitted by Mr. A.K. Singh at Indian Institute of Technology, Mumbai. This may not be quoted yet without permission.

BIBLIOGRAPHY

Census of India (1991). Population Tables, Govt. of India, New Delhi.

Kumar, A., and Tilloston, H.T. (1985). A Planning Model for Rural Roads in India. Seminar on Roads and Road Transport in Rural Areas, Organised by Central Road Research Institute, New Delhi, India

Kumar P. (1997). Facility Based Optimal Rural Road Network Design. Ph. D. Dissertation, Department of Civil Engineering, University Roorkee, India

Mahendru, A. K. Sikdar, P. K. and Khanna, S. K., “Link Efficiency in Rural (village) Road Network Planning,” Seminar on Roads and Transport in Rural Areas, CRRI, New Delhi, pp. 1-12 November 1985.

Mahendru, A. K. Sikdar, P. K. and Khanna, S. K., “Spatial Distribution and Functional planning of Settlement Hierarchies in Rural Road Networks,” Indian Highways, Journal of India Road Congress, Special Number, pp. 12-26, July 1983.

Mahendru, A. K., Sikdar, P. K. and Khanna, S. K., Nodal Points in Rural Road Network Planning, Indian Highways, Journal of Indian Road Congress, Vol. 10, pp. 5-10, 1982.

Mahendru, A. K., Sikdar, P. K. and Khanna, S. K., Some Deficiencies in Existing Practices for Rural Road Network Planning, Paper Presented at III Regional Workshop on Rural Road Development, held at Ranchi, May 1982.

Mahendru, A. K., Sikdar, P. K. and Khanna, S. K., “Linkage Pattern in Rural Road Network Planning”, Journal, Indian Roads Congress, Vol. 44-3, pp. 649-675, Dec. 1983.

Mahendru, A. K., Sikdar, P. K. and Khanna, S. K., ”Cost Consideration in Generation and Evaluation of Rural Road Network,” Journal, Indian Roads Congress, Vol. 47-3, pp. 613-643, 1986.

Mahendru, A. K., Sikdar, P. K. and Khanna, S. K., “Parameters for Generation and Evaluation of Rural Road Networks”, Proc. Of International Seminar on Rural Transportation, Indian Roads Congress, New Delhi, pp. 1-145-162, April 1989.

Mahendru, A. K., Sikdar, P. K. and Khanna, S. K., “Policy Based Rural Road Network Generation and Evaluation”, Proc. Of International Conference, ICORT-88, New Delhi, pp. 820-29, December 1988.

Mahendru, A. K., Sikdar, P. K. and Khanna, S. K., “ Rural (Village) Road Network Planning Based on Settlement Interactions”, Indian Highways, Journal Indian Roads Congress, Vol. 17(6), pp. 25-32, June 1989.

Ministry of Surface Transport (MOST) (1984). Road Development Plan for India: 1981-2001 Ministry of Surface Transport (Roads Wing), Indian Roads Congress, New Delhi, India.

Sikdar, P. K., Srivastav H. P. and Mahendru, A. K., Network Approach to Rural Road Planning, Invited Paper, Proc. National Get-Together on Road Research and its Utilisation, C.R.R.I., New Delhi, September 1992, pp. B9-B24.

Srinivasan, N. (1985). Planning of Rural Roads. Seminar on Roads and Road Transport in Rural Areas, Vol.I, Central Road Research Institute,New Delhi,Nov. 19-21, India.

Swaminathan, C.G., and Lal N.B. (1980). Appropriate Technology for Rural Road Development. Journal of Indian Roads Congress. Paper No. 329, vol. 40-4, New Delhi, India.

Swaminathan, C.G., Lal N.B. and Kumar, A. (1981). A Systems Approach to Rural Road Development Journal of Indian Roads Congress. Paper No. 346, vol. 42-4, Indian