1. Introduction

Energy is essential for economic and social development of a region or a country. However, consumption of fossil fuels is the major cause of air pollution and climate change. Improving energy efficiency and de-linking economic development from energy consumption (particularly of fossil fuels) is essential for sustainable development of a region. The energy sector, on one hand, is a part of the economy and on the other hand it itself consists of parts such as energy supply and energy demand interacting with each other. Both these interactions are of immense complexity. Energy is required for all the economic activities. Energy supplies are essential for both intermediate production as well as final consumption. So, economic development is dependent on the energy system of the country [1]. In turn, the implementation of technologies or improve¬ment of the energy system is dependent on economic factors such as capital costs, energy prices etc. Also, the demand supply balances involve the flow of energy from source as primary energy to service as useful energy. At each stage of the energy flow, technologies are involved with different conver¬sion efficiencies and losses. These complexities and inter-linkages can be understood through a model, which is a simplified representation of reality. A model is a tool for analysis, a method for clarifying the past, understanding the present, and visualizing the future.

Ecologically sound development of the region is possible when energy needs are integrated with the environmental concerns at the local and global levels. Energy planning entails preparation of area based decentralised energy plans for meeting energy needs for subsistence and development with least cost to the environment and the economy [2]. A large number of models have been developed for energy system analysis till date, and which are based on different fundamental approaches and concepts. A classification scheme will facilitate the assessment by providing in the differences and similarities between energy models. Owing to the large number of models, the classification is totally subjective. Some classify the models based only on the analytical approach, i.e. into top-down and bottom up models, whereas some classify them based on the underlying methodology into simulation, optimization, eco¬nomic equilibrium, etc. Hourcade et al. [3] uses three ways to classify energy models, viz. the purpose of the models, their structure, and their external or input assumptions. Grubb et al. [4] use six dimensions to differentiate energy models, namely top-down vs. bottom-up, time horizon, sectoral coverage, optimization vs. simulation techniques, level of aggregation, and finally geographic coverage, trade, and leakage.

As per the analytical approach, the models can be distinguished as top-down and bottom-up models (and also as hybrid models which results from the combination of the two). These two models provide quite distinct results. The differences in outcomes is due to the distinct manners in which these two types of models treat the adoption of technologies, the decision-making behaviour of economic agents, and operation of markets and economic institutions over a given period of time. The top-down models use aggregated data to examine the interactions between the energy sector and other sectors of the economy, and to examine the overall macro-economic performance of the economy. This is done by endogenising behavioural relationships. So, top-down models can examine energy policy options concerned with implications for macro-economic indicators and economy-wide emissions. In general top-down models assume that there is no discontinuity in historical patterns, i.e. historical development patterns and relationships among key underlying variables hold constant for the projection period. This assumption is not realistic in the long term, because of rapid population and economic growth. So, top-down models are suitable for predictive purposes only on the short term. They do not explicitly represent technologies and they are based on economic paradigm due to the assumption that the technol-ogy-mix results from efficient behaviour by consumer and firms under prevailing economic conditions, i.e. the most efficient technologies are given by the production frontier, which is set by market behaviour. But, the best technologies cannot be determined by market behaviour, because in developing countries predominant technologies are often based on traditional bio-fuels, which are not part of the commercial market. They consider economic sectors at highly aggregated levels and presume the economy to be in equilibrium as a result of optimal decisions taken by consumers, producers and the government. Top-down models provide greater insights into the impacts of economic policy interventions, like taxes or subsidies, which cause market distortions. These models are not realistic for developing countries because a large part of the economy is non-monetary. The macro-economic relationships ignore the influence of extensive but unreported economic activity in the informal and traditional sectors. These models also do not consider the distortions caused by non-monetary economic activities such as firewood collection and use, weak market institutions, persistent market disequilibrium, coex¬istence of numerous production functions for the same commodity and the limited choices made available to the consumers. Thus, these approaches attempt to capture the aggregate behaviour of the energy sector by use of equilibrium or partial equilibrium models which simulate prices, demand, supply and investment interactions to seek long-term market equilibrium. Such interdependent relationships are at best difficult to quantify and at worst an inaccurate description of the behaviour of markets. This approach is less suitable for economies of developing nations and smaller nations who have little control over the prices of fossil fuels and other raw materials [5].

In contrast, the bottom-up models usually focus on the energy sector exclusively, and assume interactions between the energy sector and the other sectors as negligible. So, the feedbacks from the other sectors of the economy remain exogenous to the model. These models use disaggregated data and describe energy supply processes, conversion technologies and end-use demand patterns in detail. Bottom-up models are based on optimistic engineering paradigm that provides an estimate for the technological potential by examining the effects of acquiring only the most efficient existing technol-ogies. But this theoretically predicted potential is unattainable in practice as a result of numerous social, economic (hidden costs, etc.) and legal barriers to the penetration of efficient technologies. These models can be useful in developing countries mainly because they are independent of market behaviour and production frontiers and because technologies are explicitly modelled. But, as the main drivers of the model like demand, technology change, resources remain exogenous to the model; their projections often tend to be too optimistic to be achieved by the internal savings in the economy or even including the inflow of foreign financial investment [6-8]. Hence, bottom-up or disaggregated approach of energy planning is more suitable for developing countries. With this more pragmatic approach of supply and demand, projections and investigations at disaggregated level using available data, local expertise and experiences of the energy system are possible. This approach allows important (and price-indepen-dent) effects such as technological innovations, energy transitions, market saturation and other structural shifts to be easily incorporated, which would be virtually impossible using econometric approach. Sometimes, regional macro-economic model is used as the basic framework within which demand and supply projections are made. The economies of developing countries have not reached saturation and they are yet to make most of their investment decisions. They have multiple future investment trajectories to choose from that can significantly alter their long-term technology-mix, fuel-mix and consump-tion pattern. Policies with respect to privatization, prices, taxes, trade norms, other regulatory measures, and R&D investments will have a significant impact on the consumption patterns in various end-use sectors and the competitiveness of various technologies over long run. These policies can be analyzed using top-down modelling paradigm. But, the micro-level technology and operational options available in almost every sector of the economy also offer significant scope for improvements in energy-efficiency and economic performance. Bottom-up models are useful for evaluating and implementing these short to medium-term improvement options in technol¬ogies, fuels and operational practices. For determining the long-term technology-mix, fuel-mix and resource intensity, the accumulated effects of various short-term investment decisions will be significant.

A range of methodologies is available for modelling energy systems. These include normative (optimising) models [9], system dynamic models and accounting framework simulation models. Models are also classified as simulation, optimization, econometric, macro-economic, economic equilibrium and toolbox models according to the underlying methodology. Econometric methodology use historical data (statistical techniques) to project for short term or medium term. It cannot capture structural change and does not explain determinants of energy demand, since variables are based on past behaviour, a reasonable stability of economic behaviour is required. Compared to this, the macro-economic methodology focuses on the entire economy of a society and on the interaction between the sectors. Input-output tables are used to describe transactions among economic sectors and assist in analysis of energy-economy interactions. The input-output approach can be used only when the assumptions of constant returns to scale as well as the possibility of perfect aggregation hold. These models are often developed for exploring purposes. The effects of inter-temporal preferences and long-term expectations are not taken into account, which results in a rather static representation of technical change [10].

Economic equilibrium methodologies consider the whole of the economy and focus on interrelations between the energy sector and the rest of the economy. It assumes either partial equilibrium or general market equilibrium. Partial equilibrium models focus only on equilibrium in parts of the economy, such as the equilibrium between energy demand and supply, whereas general equilibrium models are concerned with simultaneous equilibrium in all markets, as well as the determinants and properties of such an economy-wide equilibrium [11]. The disadvantage of these models is that they do not provide adequate information on the time path towards the new equilibrium, implying that transition costs are understated.

Simulation models are descriptive models based on a logical representation of a system, and they are aimed at reproducing a simplified operation of the system. A simulation model is referred to as static if it represents the operation of the system in a single time period; it is referred to as dynamic if the output of the current period is affected by evolution or expansion compared with previous periods. This model allows exploring the effects of different hypotheses via scenarios. The impacts of different assumptions and policies can be evaluated by creating different scenarios. The spreadsheet or toolbox model is always discussed as a separate methodology. It is a highly flexible model which is actually more like a software package to generate models than a model per se [12]. They often include a reference model that can be easily modified. The main disadvantage is that all important variables are indicated exogenously as parameters in future scenarios.

Optimization methodologies are used to optimize energy investment decisions or to find the least cost structure of the energy system endogenously. The outcome of an optimization model represents the best solution for given variables while meeting the given constraints. Optimization model assumes that under the given constraints perfect market condition and optimal consumer behaviour prevails. But, in developing countries, a large part of the economy is non-market based. Also a large part of the population in developing countries do not reflect consumer behaviour such as those without access to modern energy, subsistence farmers, slum dwellers, etc. An optimization model can further be distinguished by the mathematical approach used like linear programming, dynamic linear programming, non-linear programming, dynamic non¬linear programming, and mixed integer programming. Opti¬mising models typically use linear programming techniques to find out a system configuration that maximises and minimises objective functions (such as minimising costs). These have found favour in applications, such as, least cost of electricity planning and regional energy planning studies. System dynamics model makes use of engineering control theory to simulate a system as a series of interconnected stock and flow variables. System dynamics model is a powerful tool for studying the interrelationships of the different parts of a system, but their behaviour is dependent on the feedbacks between different variables, and of the relationships between those variables. Small errors in estimates of any of these values will tend to be exaggerated as the model is run for a long planning period. These models are normally not applied as practical tools for examining disaggregated energy systems. Most models of disaggregated energy systems have been based on the accounting framework approach. A set of accounting tools is provided to planners for checking the consequence and consistency of a range of scenarios. These accounting tools are simple, and provide emphasis on data, models and an easy-to-use interface.

Stockholm Environmental Institute, Boston developed a multi criteria decision support system long-range energy alternatives planning system (LEAP) to assess policy makers in evaluating energy policies and developing sound, sustainable energy plans [12]. LEAP can be used to project the energy supply and demand situation in order to glimpse future patterns, identify potential problems and assess the likely impacts of the policies. It consists of four program groups: energy scenarios, aggregation, the environmental database, and the fuel chains. LEAP serves as a database, as a forecasting tool, and as a policy analysis tool.

Members of International Energy Agency and Energy Technology Systems Analysis Program developed MARKAL (MARKet ALlocation) a large-scale model intended for the long-term analysis of energy systems at the level of a province, state, country or a region [13]. It is driven by a set of demands for energy services. MARKAL consists of a set of equations and inequations, collectively referred to as the constraints and objective functions, which is usually taken as the total discounted cost of energy system. The important feature of MARKAL is that the model allows the user to set targets and then compute an optional system's response to it.

Baumhogger et al. [14] developed MESAP (modular energy system analysis and planning) a DSS for energy and environmental management on a local, regional, or global scale. MESAP consists of a general information system based on relation database theory, which is linked to different modelling tools. It supports every phase of the structured analysis procedure to assists the decision-making process in a pragmatic way. It offers tools for demand analysis, integrated resource planning, demand side management, and simulation and optimization of supply system. In addition to this MESAP can be used to set up statistical energy and environmental information system to produce regular reports such as energy balances and emission inventories.

Voivontas et al. [15] developed a GIS-based decision support system (DSS) for evaluation of renewable energy sources (RES) potential. RES-DSS was developed in a GIS environ¬ment using MAPINFO Professional. A GIS database with data on wind, topography, urban area and special activities have been developed and used for the evaluation of theoretical potential through spatially continuous mapping of RE resources. The evaluation of wind potential is conducted in Crete island by a sequence of steps, which represents sets of restrictions on the exploitation of the potential. In RES-DSS for estimation of wind potential wind speed data necessary for estimation of theoretical potential, are modelled as region objects characterised by the attribute wind speed. Analysis presents high wind potential with wind velocity varying from 6 m/s to over 8 m/s.

Rylatt et al. [16] developed a solar energy planning system to predict and realise the potential of solar energy on an urban scale. The system will support decisions in relation to the key solar technologies: solar water heating, photovoltaics and passive solar gain. The system incorporates a domestic energy model and addresses the major problem of data collection in two ways. Firstly, it provides a comprehensive set of default values derived from a new dwelling classification scheme that builds on previous research. Secondly, GIS tools enable key data to be extracted from digital urban maps in different operational modes.

Regional integrated energy plan (RIEP) is a computer-assisted accounting and simulation tool being developed to assist policy makers and planners at district and state level in evaluating energy policies and develop ecologically sound, sustainable energy plans (Copyright Indian Institute of Science SW2659/2006). Energy availability and demand situation may be projected for various scenarios (base case scenario, high-energy intensity, and transformation, state-growth scenarios) in order to get a glimpse of future patterns and assess the likely impacts of energy policies (http://wgbis.ces.iisc.ernet.in/ energy/).

New and renewable energy policy (2005) of the Government of India (http://mnes.nic.in/Policy%20forward.htm) empha¬sises to augment energy supply to remote and deficient areas to provide normative consumption levels to all sections of the population across the country through new and renewable energy sources in furtherance of the aim of accessibility; and fuel switching through new and renewable energy system/ device deployment in furtherance of the aim of conventional energy conservation.

The regional energy planning exercise carried out for Uttara Kannada district based on RIEP involves minimisation of annual cost function to a set of equality and inequality constraints using a linear programming algorithm. The central theme of regional energy planning is the preparation of area based decentralised energy plans for meeting energy needs for subsistence and development with least cost to the environment and the economy. Centralised energy planning exercises cannot pay attention to the variations in socio-economic and ecological factors of a region which influence success of any intervention. Decentralised energy planning advocated these days is in the interest of efficient utilisation of resources, ensuring more equitable sharing of benefits from development. The regional planning mechanisms take into account all resources available and demand in a region. This implies that the assessment of the demand and supply, and the intervention in the energy system which may appear desirable due to such exercises, must be at a similar geographic scale. For example, bioresource assessment of supply and demand at the aggregate level is likely to be misleading as scarcity and surplus is always at a localised level. Consequently, the energy interventions in the form of energy supply enhancement, containing demand and/or encouraging alternative fuels may be aimed at the wrong area or target group. Normally, the district is accepted as the appropriate planning level. However, this leaves behind the complex issue of assigning boundaries for energy interventions. This is due to difference in supply and administrative boundary of a district. The geographical boundaries of the supply system in a region, based on collected biofuels, are determined by the biomass resource base and level of demand within the boundaries. In cases where fuel wood is collected from forests, the quality and accessibility of forest resource base are the most influential factors in determining the boundary of the energy system. Planned interventions to reduce energy scarcity can take various forms, such as

1. energy conservation through promotion and use of energy efficient stoves for cooking and water heating, compact fluorescent bulbs in place of ordinary incandescent bulbs,
2. supply expansions through agroforestry, farm forestry and community forestry, and
3. alternatives—renewable sources of energy such as micro/ mini/small hydropower plants and wind, solar and biomass- based systems.