|
on Computational Economics |
|
Issue of 2013‒05‒11
eleven papers chosen by |
| By: | Sancho Salcedo-Sanz (Department of Signal Theory and Communications, Universidad de Alcala, Spain.); L. Carro-Calvo (Department of Signal Theory and Communications, Universidad de Alcala, Spain.); Mercè Claramunt (Dept. Matematica Economica, Financera i Actuarial, Universitat de Barcelona, CREB, XREAP, Barcelona, Spain.); Anna Castañer (Dept. Matematica Economica, Financera i Actuarial, Universitat de Barcelona, CREB, XREAP, Barcelona, Spain.); Maite Marmol (Dept. Matematica Economica, Financera i Actuarial, Universitat de Barcelona, CREB, XREAP, Barcelona, Spain.) |
| Abstract: | Black-box optimization problems (BBOP) are dened as those optimization problems in which the objective function does not have an algebraic expression, but it is the output of a system (usually a computer program). This paper is focussed on BBOPs that arise in the eld of insurance, and more specically in reinsurance problems. In this area, the complexity of the models and assumptions considered to dene the reinsurance rules and conditions produces hard black-box optimization problems, that must be solved in order to obtain the optimal output of the reinsurance. The application of traditional optimization approaches is not possible in BBOP, so new computational paradigms must be applied to solve these problems. In this paper we show the performance of two evolutionary-based techniques (Evolutionary Programming and Particle Swarm Optimization). We provide an analysis in three BBOP in reinsurance, where the evolutionary-based approaches exhibit an excellent behaviour, nding the optimal solution within a fraction of the computational cost used by inspection or enumeration methods. |
| Keywords: | Reinsurance, Optimization Problems, Evolutionary-based algorithms |
| Date: | 2013–05 |
| URL: | http://d.repec.org/n?u=RePEc:xrp:wpaper:xreap2013-04&r=cmp |
| By: | Gianni Di Pillo (Department of Computer, Control and Management Engineering, Universita' degli Studi di Roma "La Sapienza"); Vittorio Latorre (Department of Computer, Control and Management Engineering, Universita' degli Studi di Roma "La Sapienza"); Stefano Lucidi (Department of Computer, Control and Management Engineering, Universita' degli Studi di Roma "La Sapienza"); Enrico Procacci (ACT Solutions) |
| Abstract: | This paper deals with sales forecasting in retail stores of large distribution. For severalyears statistical methods such as ARIMA and Exponential Smoothing have been usedto this aim. However the statistical methods could fail if high irregularity of sales arepresent, as happens in case of promotions, because they are not well suited to modelthe nonlinear behaviors of the sales process. In the last years new methods basedon Learning Machines are being employed for forecasting problems. These methodsrealize universal approximators of non linear functions, thus resulting more able tomodel complex nonlinear phenomena. The paper proposes an assessment of the use ofLearning Machines for sales forecasting under promotions, and a comparison with thestatistical methods, making reference to two real world cases. The learning machineshave been trained using several conguration of input attributes, to point out theimportance of a suitable inputs selection. |
| Keywords: | Learning Machines; Neural networks; Radial basis functions; Support vector machines; Sales forecasting; Promotion policies; Nonlinear optimization |
| Date: | 2013–04 |
| URL: | http://d.repec.org/n?u=RePEc:aeg:report:2013-04&r=cmp |
| By: | Peter N, Bell |
| Abstract: | This paper presents two computational techniques and shows that these techniques can improve tests for market efficiency based on profit of trading rules. The two techniques focus on interval estimates for expected profit per trade, in contrast to the standard approach that emphasizes point estimates for profit per trade (Daskalakis, 2013; Marshall, Cahan, & Cahan, 2008). The first technique uses confidence intervals to determine if the expected profit is significantly different from zero. The second technique uses moving-window resampling, a procedure of drawing sub-samples that overlap and move incrementally along a time series, to determine if the expected profit is sensitive to sample selection. The paper develops formal testing criteria based on each technique and uses simulation to establish existence results about the tests for efficiency: the standard approach can give false negative results and the new tests can give correct negative or correct positive results. Using a random walk, I show situations where the standard approach incorrectly determines that a market is inefficient whereas the new techniques do not make this error; the standard approach can be fooled by randomness of profit. Using a mean reverting process and a trading rule designed to exploit mean reversion, based on Bollinger bands, I show that the new techniques can correctly recognize an inefficient market. Since the new testing procedures can correctly identify an efficient or inefficient market, with an error rate discussed in the paper. These results support Fama’s (1970) position that trading rules can form the basis for the theory of efficient markets. This definition of an efficient market in terms of trading profit is timely given the current dominance of algorithmic trading in secondary markets. |
| Keywords: | Efficient market, trading rule, expected profit, testing procedure, confidence interval, moving window, resampling, random walk, mean reversion |
| JEL: | C63 D84 G0 G14 |
| Date: | 2013–03–15 |
| URL: | http://d.repec.org/n?u=RePEc:pra:mprapa:46701&r=cmp |
| By: | Marianna De Santis (Department of Computer, Control and Management Engineering, Universita' degli Studi di Roma "La Sapienza"); Stefano Lucidi (Department of Computer, Control and Management Engineering, Universita' degli Studi di Roma "La Sapienza"); Francesco Rinaldi (Università di Padova Dipartimento di Matematica) |
| Abstract: | Mixed-Integer optimization is a powerful tool for modeling many optimization problems arising from real-world applications. Finding a first feasible solution represents the first step for several MIP solvers. The Feasibility pump is a heuristic for finding feasible solutions to mixed integer linear problems which is effective even when dealing with hard MIP instances. In this work, we start by interpreting the Feasibility Pump as a Frank-Wolfe method applied to a nonsmooth concave merit function. Then, we define a general class of functions that can be included in the Feasibility Pump scheme for measuring solution integrality and we identify some merit functions belonging to this class. We further extend our approach by dynamically combining two different merit functions. Finally, we define a new version of the Feasibility Pump algorithm, which includes the original version of the Feasibility Pump as a special case, and we present computational results on binary MILP problems showing the effectiveness of our approach. |
| Keywords: | Mixed integer programming; Concave penalty functions; Frank-Wolfe algorithm; Feasibility Problem |
| Date: | 2013–05 |
| URL: | http://d.repec.org/n?u=RePEc:aeg:report:2013-05&r=cmp |
| By: | Marta S.R. Monteiro (Faculdade de Economia and LIAAD-INESC TEC); Dalila B.M.M. Fontes (Faculdade de Economia and LIAAD-INESC TEC); Fernando A.C.C. Fontes (Faculdade de Engenharia da Universidade do Porto and ISR-Porto) |
| Abstract: | The Hop-constrained Minimum cost Flow Spanning Tree (HMFST) problem is an extension of the Hop-Constrained Minimum Spanning Tree problem since it considers flow requirements other than unit flows. Given that we consider the total costs to be nonlinearly flow dependent with a fixed-charge component and given the combinatorial nature of this class of problems, we propose a heuristic approach to address them. The proposed approach is a hybrid metaheuristic based on Ant Colony Optimization (ACO) and on Local Search (LS). In order to test the performance of our algorithm we have solved a set of benchmark problems and compared the results obtained with the ones reported in the literature for a Multi-Population Genetic Algorithm (MPGA). We have also compared our results, regarding computational time, with those of CPLEX. Our algorithm proved to be able to find an optimum solution in more than 75% of the runs, for each problem instance solved, and was also able to improve on many results reported for the MPGA. Furthermore, for every single problem instance we were able to find a feasible solution, which was not the case for the MPGA nor for CPLEX. Regarding running times, our algorithm improves upon the computational time used by CPLEX and was always lower than that of the MPGA. |
| Keywords: | Ant Colony Optimization, Nonlinear Costs, Hybrid, Local Search, Minimum Spanning Tree Problem, Hop-constraints |
| JEL: | C61 C44 |
| Date: | 2013–05 |
| URL: | http://d.repec.org/n?u=RePEc:por:fepwps:493&r=cmp |
| By: | Gebreegziabher, Zenebe; Mekonnen, Alemu; Ferede, Tadele; Guta, Fantu; Levin, Jörgen; Köhlin, Gunnar; Alemu, Tekie; Bohlin, Lars |
| Abstract: | In response to global opportunities and domestic challenges, Ethiopia is revising its energy policy to switch from high-cost imported fossil fuel to domestically produced biofuels. Currently, there are biofuel investment activities in different parts of the country to produce ethanol and biodiesel. However, there is no rigorous empirical study to assess impacts of such investments. This paper assesses the distributive effect and food security implications of biofuels investment in Ethiopia, using data from 15 biofuels firms and 2 NGOs in a CGE (computable general equilibrium) analysis. Findings suggest that biofuels investments in the context of Ethiopia might have a ‘win-win’ outcome that can improve smallholder productivity (food security) and increase household welfare. In particular, the spillover effects of certain biofuels can increase the production of food cereals (with the effect being variable across regions) without increasing cereal prices. When spillover effects are considered, biofuel investment tends to improve the welfare of most rural poor households. Urban households benefit from returns to labor under some scenarios. These findings assume that continued government investment in roads allows biofuels production to expand on land that is currently unutilized, so that smallholders do not lose land. Investment in infrastructure such as roads can thus maximize the benefits of biofuels investment. |
| Keywords: | biofuels investment, CGE model, food security, household welfare, equivalent variation, Ethiopia |
| JEL: | Q56 Q42 O44 O5 |
| Date: | 2013–01–18 |
| URL: | http://d.repec.org/n?u=RePEc:rff:dpaper:dp-13-02-efd&r=cmp |
| By: | Gianluca Amori (Department of Computer, Control and Management Engineering, Universita' degli Studi di Roma "La Sapienza"); Luca Becchetti (Department of Computer, Control and Management Engineering, Universita' degli Studi di Roma "La Sapienza"); Giuseppe Persiano (Università di Salerno Dipartimento di Informatica); Andrea Vitaletti (Department of Computer, Control and Management Engineering, Universita' degli Studi di Roma "La Sapienza") |
| Abstract: | With the growing di usion of devices with wireless communication capa- bilities as well as the success of social networking platforms, it has become more and more interesting to study dynamic evolving networks, in which the set of connections (however de ned) between the agents in the network varies over time. For instance, ad-hoc mobile networks are evolving networks in which a number of mobile hosts are free to move about a given area and capable when close enough of interacting with each other over wireless links without the need of an underlying backbone network. Other examples include P2P networks and in general social network contexts, in which the users dynamically establish and terminate social interactions. The topology of such networks changes over time, as edges (either directed or undirected) that represent interactions between nodes are dynamically added or removed in the network graph. Computing measures of centrality in such scenarios can be a challenging task. Classic measures of centrality and nodes' importance from graph theory and network analysis can be computed by a centralized en- tity on an aggregated representation of a dynamic network. However, privacy and/or scalability issues, or simply the absence of central coordination, may suggest a fully decentralized approach in which the computation is carried out by each node considering its own interactions with other nodes in the net- work. In this Master's thesis we propose lightweight algorithms for computing some importance measures of nodes in a dynamic evolving network in a fully decentralized way, without any knowledge of the whole network structure or assumptions on its future evolution. In particular, the main part of our work regards the computation of decentralized estimations of Google's PageRank and its theoretical analysis as a problem of random walks on dynamic graphs. We also introduce algorithms for computing some classical degree centrality measures with e cient use of resources. As it turns out, while straightfor- ward in a centralized setting, some of these measures are hard to compute in a fully decentralized way. We analyze all these algorithms in terms of hardware resources (storage space and computational power) required at each node as well as time complexity and network overhead in the transmissions, showing how they are implementable also on low-power devices such as RFID tags. We also run simulations of our algorithms on real-world dynamic evolving network data and show their performances with respect to centralized computations of analogous measures on aggregated static representations of such networks. |
| Keywords: | Dynamic Networs; importance measures; PageRank |
| Date: | 2013–06 |
| URL: | http://d.repec.org/n?u=RePEc:aeg:report:2013-06&r=cmp |
| By: | Jeuland, Marc; Whittington, Dale |
| Abstract: | This article develops a “real options” approach for planning new water resources infrastructure investments and their operating strategies in a world of climate change uncertainty. The approach is illustrated with an example: investments in large new multipurpose dam alternatives along the Blue Nile in Ethiopia. The approach incorporates flexibility in design and operating decisions – the selection, sizing, and sequencing of new dams, and reservoir operating rules. The analysis relies on a simulation model that includes linkages between climate change and system hydrology, and tests the sensitivity of the economic outcomes of investments in new dams to climate change and other uncertainties. Not surprisingly, the results for the Blue Nile basin show that there is no single investment plan that performs best across a range of plausible future climate conditions. The value of the real options framework is that it can be used to identify dam configurations that are both robust to poor outcomes and sufficiently flexible to capture high upside benefits if favorable future climate and hydrological conditions arise. The real options approach could be extended to explore design and operating features of development and adaptation projects other than dams. |
| Keywords: | ile Basin, real options, dams, climate adaptation, cost-benefit analysis, Ethiopia, Monte Carlo simulation |
| JEL: | D81 O22 Q25 Q42 Q54 |
| Date: | 2013–03–06 |
| URL: | http://d.repec.org/n?u=RePEc:rff:dpaper:dp-13-05-efd&r=cmp |
| By: | Marco Patriarca; Anirban Chakraborti |
| Abstract: | We discuss several multi-agent models that have their origin in the kinetic exchange theory of statistical mechanics and have been recently applied to a variety of problems in the social sciences. This class of models can be easily adapted for simulations in areas other than physics, such as the modeling of income and wealth distributions in economics and opinion dynamics in sociology. |
| Date: | 2013–05 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:1305.0768&r=cmp |
| By: | Nazarian, Rafik; Gandali Alikhani, Nadiya; Naderi, Esmaeil; Amiri, Ashkan |
| Abstract: | Recently, with the development of financial markets and due to the importance of these markets and their close relationship with other macroeconomic variables, using advanced mathematical models with complicated structures for forecasting these markets has become very popular. Besides, neural network models have gained a special position compared to other advanced models due to their high accuracy in forecasting different variables. Therefore, the main purpose of this study was to forecast the volatilities of TSE index by regressive models with long memory feature, feed forward neural network and hybrid models (based on forecast combination approach) using daily data. The results were indicative of the fact that based on the criteria for assessing forecasting error, i.e., MSE and RMSE, although forecasting errors of the feed forward neural network model were less than ARFIMA-FIGARCH model, the accuracy of the hybrid model of neural network and best GARCH was higher than each one of these models. |
| Keywords: | Stock Return, Long Memory, Neural Network, Hybrid Models. |
| JEL: | C14 C22 C45 C53 |
| Date: | 2013–03–15 |
| URL: | http://d.repec.org/n?u=RePEc:pra:mprapa:46786&r=cmp |
| By: | Richard S.J. Tol (Department of Economics, University of Sussex, UK; Institute for Environmental Studies, Department of Spatial Economics, Vrije Universiteit, Amsterdam, The Netherlands); Muireann Lynch (Electricity Research Centre, University College, Dublin, Ireland); Aonghus Shortt (Electricity Research Centre, University College, Dublin, Ireland); Mark O’Malley (Electricity Research Centre, University College, Dublin, Ireland) |
| Abstract: | We employ Monte Carlo analysis to determine the distribution of returns for various electricity generation technologies. Costs and revenues for each technology are arrived by means of a sophisticated unit commitment and economic dispatch algorithm. The results show that small amounts of coal investment along with high investment in advanced CCGT can reduce the risk of baseload-only portfolios, while flexible generation technologies appear on the efficient frontier when all technology types are considered. Diversification incentives regarding operational considerations dominate over incentives to diversify between fuel types |
| Keywords: | Power generation, mean-variance portfolio |
| JEL: | Q40 |
| Date: | 2012–10 |
| URL: | http://d.repec.org/n?u=RePEc:sus:susewp:4012&r=cmp |