 Thank you. Thank you very much. Actually, I'm very delighted to be here. The problem is, I don't want to come back and talk to you. So, to be an ASM of a great owner, I spent two years here and that was an amazing time. And I get very good of those keywords and expertise that helped me to overcome the development I had in London. So, the transition was very smooth. Nothing serious. When I went to London, I thought that I was registered in business school and I will do my research in business school. But actually, I was told that I'm going into discipline at work, so I have to be not only in the business school but also in the computer science department. So, three supervisors were assigned. So, I have three supervisors and it's very hard to be supervised by three people. Three. Actually, two of them are, one is mathematician, one is expert in computational optimization and the other one is an economist, especially interested in energy policy. So, we were talking with Mr. Arjan, and I was saying that doing PhD at Imperial College, which is well known, I mean, the best in interdisciplinary research around the UK and I guess in Europe also. So, it's very tough because I have to do the other side because I have to provide for the department. So, yeah. So, we slightly changed the topic in the title because this gives more information not to be most specific, but I will talk about renewable energy and we'll skip to uncertainty that is prevalent in the UK, which is the market. We'll talk about optimization models and energy models. But myself, I'm more interested in wind energy because I'm doing research on wind energy and specifically I'm doing research on artillery because the UK has the most capacity in Europe in offshore wind and more and more projects are being built. So, the overall portfolio is very attractive for investors when there is much offshore wind. So, wind has some problems. For example, load factors, you know, sometimes wind blows, sometimes it doesn't blow and there is also whether the turbines are on shore or in offshore, in the ocean or in the sea. And wind turbine, we can see from the power curve when wind speed is high, then we get much load factor. There are more than four gigawatts of offshore wind capacity in the European Union and more than 50% capacity is told in the UK and as I said already, many projects are underway but it has high investment costs, low variable costs, basically it has no fuel costs and the major problem is the intermittency. And this will grow when in the future there will be more offshore wind penetration and we get electricity market and we should deal with it. The impact of intermittency is a short-term balancing because the system operator has to match demand and supply. So, it's a huge problem for system operator. There is a lot of wind. So, high level wind will raise this. And the estimated cost of intermittency is between 501 million pounds. I mean, the UK electricity and energy policy always looks ahead and plans what energy market will look like in 2020, 2030 and based on the estimates of whether the clients go right or wrong. What does that mean, the estimated cost of intermittency? Is the cost of the sort of revenue not collected? Yes, I mean, if you cannot serve the demand what's the impact of that? Because, for example, system operator has to give some electricity, for example, we've got a lot of electricity. So, the replacement cost of the energy not created? Yes, there are a lot. You know, the wind provides energy equal to 30% of its capacity but the relay of the capacity is 10%. But it has some spare capacity, which is a lot. So, it's an energy flow for the UK. We can see that oil is mainly used in transport, or transport and heat, also gas, gas for heat and renewable power, basically goes with electricity. And this, you can see that it's a renewable flow of renewable sources. Basically what this shows is that either it is lost or it goes to the electricity. So, there are different types of renewable sources and this is a generation very frequent for a particular time, for example, in November 2011. And we can see that this is, okay, there are three types of generation plans, a generating plan. The first one are called base load plans. Basically they are nuclear and all the plans that have low wearable costs, but of course the high investment costs. And mid-load plans, which are combined cycle gas turbines and sometimes coal, but coal is more going to the peak. And this is a generation mix for the UK. We can see that it imports also electricity from France and the red goes to hydrogen. So this is a picture for the UK and if we look at the future of wind power, we will have more surface in this. Each generation will get into interviews? Yeah, generation. What does that mean in France? France, it means whatever is left, it imports from France. The difference in character with that test. Because in Europe, for example, the UK is connected to Sweden, to Denmark. Whenever it needs some electricity, it has some electricity shortages. It can just online, it can ask and then import electricity. But it's a very high development. They have very high development rates and they are very connected. Vitaly, each tooth is one day, right? Measure in half hour from midnight on one moment. The whole thing is one day, right? No. Which one? Yes, it is. Is this what it is showing? The electricity consumption for one day measured in half hour. It is from long to some amount, right? Oh, okay, for one week. One week, right? We can see the variations. So estimated UK renewable capacity in 2020. We can see a lot of wind, onshore, offshore. Some goes to tidal, some to wave and some other sources. So this is kind of benchmark for policy makers to achieve that. UK targeted that in 2020. I won't speak about that. In 2020, 20% of all energy should come from renewable sources. But in 2030, they targeted that 80% should come from renewable sources. So, but looking at these steps, how they go, it doesn't sound plausible. And my name, not now, will go to the investment patients and uncertainties that are in this market. Is to show that investors can basically improve their investment strategies when they invest heavily in offshore wind. That's one of my aim, one of my research. So electricity policy plays a huge role in the development of other countries. But because it has many schemes, for example, it has decarbonisation policy. They call decarbonisation policy. And basically what it does, create some tools for generators to use them to get some subsidies. For example, renewable education certificates, they are green certificates. They are called renewable rocks in the UK. Which basically certificates that are traded and like costs are split from consumers to producers and suppliers. And depending on the situation. And there's a governing body of them that governs all these circulation of rocks. So is it like carbon credits? Kind of yes, carbon credits. But more carbon credits because there are two types of billion turrets now. Billion turrets contract for differences. It's like you sign a contract for 20 years. And you agree to get, for example, eight tries, whatever you sell. And depending on what price you sell in the spot market, you get the differences topped up. For example, if you agree at 70, right, price. If you sell a spot price 120, you have to give back the remaining part. But if you sell less than the system, whatever governing body it is will give you the rest. This is called billion turrets in contract for difference. It means a fixed price. But if it fails, then there's a premium in billion turrets. Which means whatever you sell in the spot market, you will get some extra bonus. For example, 20% or 30%. And it also will fluctuate like in the spot market. And the rest comes from European policy. For example, carbon price law. They don't have carbon price law, but they took from European policy. And they just added some number and they increased for UK. For example, you should pay for carbon a 15.7 pound per ton CO2 emitted in 2013. If it's straight line, it goes up and reaches 30 pounds in 2020. And it goes even higher. And there is an ambitious performance standard. You cannot emit more than 450 ground CO2 per kilo of power produced. Basically, bands and construction and no fuel stations with a lot of capacity. Is it an average standard? It would be hard to keep this standard. You cannot produce more than this. We should keep something else. What? If you use coal for power generation, should you keep this standard? Yes. It would be harder. The aim of the policy is to reduce this kind of plant. This is a physical output when you burn coal. How can you wait? The system has to come to that level. This is what I say. Is it average? It's the system average, not the plant average. Yes, this is the system average. It's not a plant. And also, there is a capacity mechanism which basically deals with situative supplies because they facilitate the carbonization policy, the power going to secure affordable, cheap, meeting all the requirements, all these collections that are discussed in the capacity mechanism. What I'm interested in is more about the firms that want to invest in electricity generation. I said 20%, but actually 15% should come from renewables in 2020. It basically means energy will come from renewables. It basically means that more than 15% of electricity will come from renewables. Without big hydro probably? Yeah. Each country in the EU has adopted their own guidelines for achieving by 2020 or 2050 or whatever. The most tricky is UK. If we compare, they set higher standards for them. And they want to be the warrants. They're not that high. There are countries in the EU that have set much higher goals for them for 2020. 15% is, I think, average for EU. Yeah. For 2020, I agree, but for 2030, it goes 80%. They want to, I don't know, what's the logic behind it? No, Russia. It's a policy against Russia. Russian gas. Investments are made by big companies, the British gas, EGF Energy, EONPK, and the big six. The problem is that if they go through this decarbonisation policy, that would create problems for these big companies and they would not be interested in investing. And basically, they looked at all the aspects of their policy and they want to attract more medium-sized and small-sized companies to come and invest in UK electricity. So, I don't know, this is kind of very interesting, very demanding. We know that investment, there are some strategies in investment decision making. And I'm looking at how to basically to model this investment in electricity market. Basically, there are two types, more types. Basically, if you look at uncertainties, then you will deal with stochastic models. And if it's dynamic, then it means dynamic and stochastic, because real options basically use dynamic programming or stochastic programming. But if you use your static MPV, it's deterministic and static, and it has some drawbacks, although MPV is just one lump sum of investment and you just invest and wait a period and get some reward. And, on the contrary, in real options, you deal with incoming information and then you update your investment policy each time. So, this is kind of introduction of what the investment strategies are. And particularly, there are two types of models, short-term and long-term models in investment in electricity. Short-term models are basically unique models. Basically, it means what you want to build and operate. And they are solved, these kinds of problems are solved with dynamic programming, mixed-integer programming, maybe I should say what the unique commitment means. For example, if there are different units of energy sources and your job is to determine which plants should operate which do not. So, this is a problem, and if you bring it to the mathematical form, they become, for example, mixed-integer programming problems, which are very hard to solve, and sometimes they become intractable. That's why some heuristics are used, for example, that grantion and relaxation is used to facilitate this problem. But basically, these kinds of unique commitment problems suffer from this curative dimensionality, and sometimes there are no algorithms to solve it. The other one I call economic regeneration, these types of models, is basically it comes up to unique commitment problems when you already specify which kind of root sources should operate and then how much they should serve to the system operator. Basically, there are linear programming models. If number of constraints and number of unknowns are very large, so you can use heuristics. For example, it's made of this top, this one of them, which basically ranks plants or sources based on the variable process. The first one is, for example, nuclear, biomass, when they come to gas, coal, and then, again, sources. The long-term models are passage management problems, and it's like dealing with through time how much capacity should be added. The capacity expansion is not for electricity, but to generate capacity expansion. It means that you are dealing with time and scale, but for particularly in the electricity market, it means that you should solve multi-stage stochastic problems or you can, depending on the model, you can use stochastic optimum control, but this excludes the facilities location problem, because if we include facilities locations, it becomes very hard to solve. That's why it's sometimes complicated. Time and scale are very important in the geological expansion. So, what are the uncertainties? 10 uncertainties, but they can be more. The first one, electricity demand level, it is projected that it will rise, and it's very hard to, yeah, to rise, but how much it will rise, and what the pace will be in there. That creates uncertainty for investors. Full price is another source of uncertainty. Gas oil, et cetera, because some prices are correlated with gas and oil. Electricity spot prices, because electricity is traded in spot prices, or if you have some bilateral contracts, you are kind of secured from these spot prices. Carbon prices will rise, I mean, fall, but that's kind of uncertain. Climate, of course, climate fluctuations, for example, rainfalls. And storage is, there's a huge research going on on storage in the UK, in Europe, in the US also. And now it is considered not viable. However, there are some new technologies that can store electricity that are so negligible and cost so high that they are practically viable. So storage and holding costs, if it's viable, then that creates uncertainty because the policy change in our issue is particularly for the UK. It's a huge uncertainty. Technology for change, I think it's for any sphere, for any industry, this is uncertain. And also the investment costs. For example, for offshore wind, investment costs rise, and it has some objective and subjective reasons why it rises, because new technologies arrive and offshore distance from shore defends and new turbines are built in deeper places and that's why it costs rise. But there's some trade-off, whether it's useful to have these kind of turbines because investors don't want to spend much money on the offshore wind. Vitaly, sorry. Going back to the, if you have the electricity demand level being an uncertainty, why don't you have the interventancy in this list? Interventancy. Climate fluctuations probably represent... We can include in this. But that's a very unique... More specific to maybe... No, sun hasn't... Hydro has interventancy. But it's for portfolio, I mean, for energy. Right? Because fuel prices concerns to every source. Carbon prices... Basically covers almost everything that's why... Maybe I should have included it, but... It's actually a unique characteristic of renewable energies, the fact that they're so unstable as a source. Yeah, hydrogen. Some of them aren't. Except hydrogen. Hydro, geothermal... Yeah. Ultimately, what I want to create in my research is to create a model which will best describe UK investment atmosphere. Right? There are different models. For example, price, political content, it's an energy system model that was first implemented in 1970. In the US. NANDs, for example, also in the US. But the problem is that these models are simple. They use linear problem approach. Some use heuristics, for example. They use some algorithm, which is an interactive algorithm. And some use simulation. But my aim is... I want to create a long-term investment model for the UK electricity industry. And each will be designed with the help of stochastic programming, maybe dynamic programming or dynamic stochastic programming, and will have some game theoretical aspects. Because of these renewable certificates, because of these many interactions in the market, that will create some game theoretical aspects. This is the ultimate aim. Is that the only argument or why you decided to use this kind of methodology? Because there is no such model in the UK now. And recently, because it's ongoing process, they are formulating their policy, and there are some research, but nobody is just... Actually, this dynamic, this stochastic model here, is... Yeah, it's not governmental, but... They were initiators for this model. This incorporates this past expansion problem, also this economic and generational dispatch model. But I looked their website, and this model, this is private one, that's why it's a question mark here. But when I looked the description, they said they are going to solve problems dealing with uncertainty, using uncertain optimization problems. Basically, what they do is using simulation. They call it uncertain optimization on the uncertainty. It's more a private product. It's not academy, but professional model. It's going to cost you if you decide to do this. That's why the argument is that we want to have an academy model which will deal with all kinds of uncertainties, prescribe uncertainties, and then we'll have some huge impact, hopefully. My team, I'm seeing supervisors. Richard Wayne is a key figure in energy policy in the professor of sustainable energy business at Imperial. Burke Raston is a professor of computational methods, of computer science, and Dr. Parnas Barker is also a professor of computational optimization. And this is my research title at Imperial. Optimization will be investment decision-making for an energy portfolio that will be enlarging on the offshore region. Basically, I want to convince investors that our ultimate goal is to set portfolios that will have more offshore wind and also into an obvious... After that, you get a very comfortable job at the energy company. Yeah, also from industry. Not only from academia. At what stage are you right now? This is my first year. I did a lot of research and literature review and came up with my ideas and proposed, and I don't know... They said that some are good, some are bad, some are more ambitious. But finally, I could push and say, I want to create a model for UK because, yeah, I know that there is a dynamic dispatch model, but it's not professional, and I want to create some, and maybe they will be interested in the model. And they said, yeah, that's a challenging, but it can be. In your research of Jenny... What are the factors that distributed in an energy generation factor? How important it is for decision-making because we know that a lot of energy is lost through transmission, 20% or more, so maybe I missed something, but I didn't see that factor being, like, important one. Yeah. Because I'm sorry, why I said, because you brought in multiples of massive renewable energy production, like offshore, which goes the same path as traditional conventional energy generation system does. They are focused constantly in one place and then they distribute all over the country. So why this factor is not taking... Maybe I missed it. No, it's a more engineering issue, right? Because how the system is... Decision-making as well. It's not exactly... Let me elaborate on that, if you don't mind. See, in the world now, we see a lot of tendencies as far as the renewable energy is concerned, favoring distributed generation on-site. So I don't know what's the situation in the UK, but you might want to consider that factor in your model, meaning that there are so many factors you mentioned. So maybe in the UK, the general policy starts favoring distributed generation, which will affect offshore wind, right? So what we suggest, I guess, is think about it. Maybe you want to consider this factor in your model. Yeah, sure. There can be more aspects that I can maybe omit it, because it's... so much is going on and it's the first year and I'm just looking for new... Thank you. Of course the EU is moving towards total liberalization of electricity and gas markets and right now there are even possibilities for distributed... distribution system to be so liberalized that some companies just own infrastructure and they charge specific rate for moving electrons from point A to point B. So retailers can just resell it. So in your opinion, what works more efficiently this way? I mean, the way that EU is moving or vertically integrated companies that it's just easier to estimate specific costs and to keep them based on public utilities, regulatory commission's decisions to charge specific amount and then somehow to regulate the prices that otherwise in the liberalized market might go... Personally, I'm prone to the liberalized market because, yeah, there can be mass when you liberalize the market. Many first will enter who knows what's going on. But when you have a huge, great country which talented people who design a policy and they know what they actually want, I've read many policy papers and they deal with every aspect of each type of generation. For example, there are plants, as you said coal plants, right? Or they can ban production of coal plants of 450. But you can sequester the carbon, capture the carbon and so on. There's a carbon capture and sequestration model and they already discussed it and many just consider this as an option. If you build a plant, as you see that you are not building the limits, you can upgrade your plant and then, yeah, you can upgrade. I mean, everything is designed in detail but of course there are some publishers. Again, you mentioned in the factors the demand side and what will be the role of energy savings and energy efficiency in this demand because demand might go in two ways. It could be larger or it could be smaller due to this policy. Parallel to this policy, this decarbonisation policy, they also promote energy savings and there are some local companies who offer some consultancy and they go door to door and they say, ok, if you unplug your iron, gold, isolation, it's very highly developed and most of the households do. They hire this consultancy and they do that. Recently, also there are some projects, for example, they say, ok, we'll bring solar panels and install on the roof of your house and you don't have to pay for your electricity. Whatever you produce, that will go to the grid and they will distribute this but you don't have to pay anything. This, that you create this electricity, that all of it pays for you. The others who don't have that, they pay for it. These kind of sophisticated models exist in the UK. Nothing? I think that's great. Thank you. There is a number of models in the market. And in these models, there is some type of dynamic system, differential or stochastic differential equation. I would like to ask what is your new idea to improve the existence of stochastic models? There are models, generally, about, I mean, for wind energy, there is a lot of research and they use stochastic models but for UK, I mean for the UK, they don't go deep to use these stochastic models and so on. That's why my supervisors are one of the experts in this field and they say there are models a lot but the problem is that for the UK, they don't use stochastic programming, for example. They don't use stochastic dynamic programming. They don't use robust programming and they are experts in robust programming. They want to create a system which will be robust. I mean, whatever policy, it will be robust to their inputs. Eventually they want to achieve that. Now I'm not in the stage to decide which kind of algorithm I'm going to use but now I'm more getting acquainted with research and literature review and seeing what the gaps are and I want to design a model for UK. That's my advantage because nobody has done but there are tons of research in particular for me. Surprises are stochastic random. They simulate. Mostly they simulate. Since for differential equations stochastic, we cannot write explicit solutions. This is a problem. That's why they simulate. We get some results and say we have satisfied with this. We simulate up to 2020 or 2015 and they say that's the scenario. Commercial companies like Michel, Siemens, they do their own research but we don't have access to this kind of research. Maybe they have a model but I'm going to mention that. Let's go back to the list of our circles. For example, the policy changes and the technology changes. Do we have an idea how it will be modulated? I struggled about this. For example, the policy change. It can be modelled like you can generate three or four types of possible policies that can be in the future. And then in your model, you can model this like integer variables. If it does model occurs, then this comes to your model. This is like one, I mean, the sum you can write. If one operates, that would be one. The others would be zero. So as I understand, modelling will be based on some assumption and you will run some scenario based on the assumptions. You cannot predict which policies will be. Yeah, you cannot predict. You can't. I mean, you have some candidates. For example, in policy, for example, two types of schemes for contracts. You can have fidget tariffs or you can have premium tariffs. You don't know which one will ultimately be coming. So you can have model one and then if it fails, just switch to the other one. Or you can modelled two different models with two different scenarios with this.