Avaluació de la tecnologia de les GPUs

En el projecte s’ha dut a terme un estudi sobre la tecnologia que aporten les targetes gràfiques (GPU) dins l’àmbit de programació d’aplicacions que tradicionalment eren executades en la CPU o altrament conegut com a GPGPU. S’ha fet una anàlisi profunda del marc tecnològic actual explicant part del maquinari de les targetes gràfiques i de què tracta el GPGPU. També s’han estudiat les diferents opcions que existeixen per poder realitzar els tests de rendiment que permetran avaluar el programari, quin programari està dissenyat per ser executat amb aquesta tecnologia i quin és el procediment a seguir per poder utilitzar-los. S’han efectuat diverses proves per avaluar el rendiment de programari dissenyat o compatible d’executar en la GPU, realitzant taules comparatives amb els temps de còmput. Un cop finalitzades les diferents proves del programari, es pot concloure que no tota aplicació processada en la GPU aporta un benefici. Per poder veure millores és necessari que l’aplicació reuneixi una sèrie de requisits com que disposi d’un elevat nombre d’operacions que es puguin realitzar en paral lel, que no existeixin condicionants per a l’execució de les operacions i que sigui un procés amb càlcul aritmètic intensiu.

A New Model for Designing a Product Line Using TURF Analysis

This paper introduces the approach of using TURF analysis to design a product line through a binary linear programming model. This improves the efficiency of the search for the solution to the problem compared to the algorithms that have been used to date. Furthermore, the proposed technique enables the model to be improved in order to overcome the main drawbacks presented by TURF analysis in practice.

The achievable region approach to the optimal control of stochastic systems

The achievable region approach seeks solutions to stochastic optimisation problems by: (i) characterising the space of all possible performances(the achievable region) of the system of interest, and (ii) optimisingthe overall system-wide performance objective over this space. This isradically different from conventional formulations based on dynamicprogramming. The approach is explained with reference to a simpletwo-class queueing system. Powerful new methodologies due to the authorsand co-workers are deployed to analyse a general multiclass queueingsystem with parallel servers and then to develop an approach to optimalload distribution across a network of interconnected stations. Finally,the approach is used for the first time to analyse a class of intensitycontrol problems.

Network revenue management with product-specific no-shows

Revenue management practices often include overbooking capacity to account for customerswho make reservations but do not show up. In this paper, we consider the network revenuemanagement problem with no-shows and overbooking, where the show-up probabilities are specificto each product. No-show rates differ significantly by product (for instance, each itinerary andfare combination for an airline) as sale restrictions and the demand characteristics vary byproduct. However, models that consider no-show rates by each individual product are difficultto handle as the state-space in dynamic programming formulations (or the variable space inapproximations) increases significantly. In this paper, we propose a randomized linear program tojointly make the capacity control and overbooking decisions with product-specific no-shows. Weestablish that our formulation gives an upper bound on the optimal expected total profit andour upper bound is tighter than a deterministic linear programming upper bound that appearsin the existing literature. Furthermore, we show that our upper bound is asymptotically tightin a regime where the leg capacities and the expected demand is scaled linearly with the samerate. We also describe how the randomized linear program can be used to obtain a bid price controlpolicy. Computational experiments indicate that our approach is quite fast, able to scale to industrialproblems and can provide significant improvements over standard benchmarks.

Parallel scheduling of multiclass M/M/m queues: Approximate and heavy-traffic optimization of achievable performance

We address the problem of scheduling a multiclass $M/M/m$ queue with Bernoulli feedback on $m$ parallel servers to minimize time-average linear holding costs. We analyze the performance of a heuristic priority-index rule, which extends Klimov's optimal solution to the single-server case: servers select preemptively customers with larger Klimov indices. We present closed-form suboptimality bounds (approximate optimality) for Klimov's rule, which imply that its suboptimality gap is uniformly bounded above with respect to (i) external arrival rates, as long as they stay within system capacity;and (ii) the number of servers. It follows that its relativesuboptimality gap vanishes in a heavy-traffic limit, as external arrival rates approach system capacity (heavy-traffic optimality). We obtain simpler expressions for the special no-feedback case, where the heuristic reduces to the classical $c \mu$ rule. Our analysis is based on comparing the expected cost of Klimov's ruleto the value of a strong linear programming (LP) relaxation of the system's region of achievable performance of mean queue lengths. In order to obtain this relaxation, we derive and exploit a new set ofwork decomposition laws for the parallel-server system. We further report on the results of a computational study on the quality of the $c \mu$ rule for parallel scheduling.

Restless bandits, partial conservation laws and indexability

We show that if performance measures in a stochastic scheduling problem satisfy a set of so-called partial conservation laws (PCL), which extend previously studied generalized conservation laws (GCL), then the problem is solved optimally by a priority-index policy for an appropriate range of linear performance objectives, where the optimal indices are computed by a one-pass adaptive-greedy algorithm, based on Klimov's. We further apply this framework to investigate the indexability property of restless bandits introduced by Whittle, obtaining the following results: (1) we identify a class of restless bandits (PCL-indexable) which are indexable; membership in this class is tested through a single run of the adaptive-greedy algorithm, which also computes the Whittle indices when the test is positive; this provides a tractable sufficient condition for indexability; (2) we further indentify the class of GCL-indexable bandits, which includes classical bandits, having the property that they are indexable under any linear reward objective. The analysis is based on the so-called achievable region method, as the results follow fromnew linear programming formulations for the problems investigated.

Measuring and explaining farm inefficiency in a panel data set of mixed farms

This paper aims to estimate a translog stochastic frontier production function in the analysis of a panel of 150 mixed Catalan farms in the period 1989-1993, in order to attempt to measure and explain variation in technical inefficiency scores with a one-stage approach. The model uses gross value added as the output aggregate measure. Total employment, fixed capital, current assets, specific costs and overhead costs are introduced into the model as inputs. Stochasticfrontier estimates are compared with those obtained using a linear programming method using a two-stage approach. The specification of the translog stochastic frontier model appears as an appropriate representation of the data, technical change was rejected and the technical inefficiency effects were statistically significant. The mean technical efficiency in the period analyzed was estimated to be 64.0%. Farm inefficiency levels were found significantly at 5%level and positively correlated with the number of economic size units.

On bounds for network revenue management

The Network Revenue Management problem can be formulated as a stochastic dynamic programming problem (DP or the\optimal" solution V *) whose exact solution is computationally intractable. Consequently, a number of heuristics have been proposed in the literature, the most popular of which are the deterministic linear programming (DLP) model, and a simulation based method, the randomized linear programming (RLP) model. Both methods give upper bounds on the optimal solution value (DLP and PHLP respectively). These bounds are used to provide control values that can be used in practice to make accept/deny decisions for booking requests. Recently Adelman [1] and Topaloglu [18] have proposed alternate upper bounds, the affine relaxation (AR) bound and the Lagrangian relaxation (LR) bound respectively, and showed that their bounds are tighter than the DLP bound. Tight bounds are of great interest as it appears from empirical studies and practical experience that models that give tighter bounds also lead to better controls (better in the sense that they lead to more revenue). In this paper we give tightened versions of three bounds, calling themsAR (strong Affine Relaxation), sLR (strong Lagrangian Relaxation) and sPHLP (strong Perfect Hindsight LP), and show relations between them. Speciffically, we show that the sPHLP bound is tighter than sLR bound and sAR bound is tighter than the LR bound. The techniques for deriving the sLR and sPHLP bounds can potentially be applied to other instances of weakly-coupled dynamic programming.

Optimization of multiclass queueing networks with changeover times via the achievable region method: Part II, the multi-station case

We address the problem of scheduling a multi-station multiclassqueueing network (MQNET) with server changeover times to minimizesteady-state mean job holding costs. We present new lower boundson the best achievable cost that emerge as the values ofmathematical programming problems (linear, semidefinite, andconvex) over relaxed formulations of the system's achievableperformance region. The constraints on achievable performancedefining these formulations are obtained by formulatingsystem's equilibrium relations. Our contributions include: (1) aflow conservation interpretation and closed formulae for theconstraints previously derived by the potential function method;(2) new work decomposition laws for MQNETs; (3) new constraints(linear, convex, and semidefinite) on the performance region offirst and second moments of queue lengths for MQNETs; (4) a fastbound for a MQNET with N customer classes computed in N steps; (5)two heuristic scheduling policies: a priority-index policy, anda policy extracted from the solution of a linear programmingrelaxation.

Looking for orders of magnitude : five essays in trade and the environment and economic geography

General Summary Although the chapters of this thesis address a variety of issues, the principal aim is common: test economic ideas in an international economic context. The intention has been to supply empirical findings using the largest suitable data sets and making use of the most appropriate empirical techniques. This thesis can roughly be divided into two parts: the first one, corresponding to the first two chapters, investigates the link between trade and the environment, the second one, the last three chapters, is related to economic geography issues. Environmental problems are omnipresent in the daily press nowadays and one of the arguments put forward is that globalisation causes severe environmental problems through the reallocation of investments and production to countries with less stringent environmental regulations. A measure of the amplitude of this undesirable effect is provided in the first part. The third and the fourth chapters explore the productivity effects of agglomeration. The computed spillover effects between different sectors indicate how cluster-formation might be productivity enhancing. The last chapter is not about how to better understand the world but how to measure it and it was just a great pleasure to work on it. "The Economist" writes every week about the impressive population and economic growth observed in China and India, and everybody agrees that the world's center of gravity has shifted. But by how much and how fast did it shift? An answer is given in the last part, which proposes a global measure for the location of world production and allows to visualize our results in Google Earth. A short summary of each of the five chapters is provided below. The first chapter, entitled "Unraveling the World-Wide Pollution-Haven Effect" investigates the relative strength of the pollution haven effect (PH, comparative advantage in dirty products due to differences in environmental regulation) and the factor endowment effect (FE, comparative advantage in dirty, capital intensive products due to differences in endowments). We compute the pollution content of imports using the IPPS coefficients (for three pollutants, namely biological oxygen demand, sulphur dioxide and toxic pollution intensity for all manufacturing sectors) provided by the World Bank and use a gravity-type framework to isolate the two above mentioned effects. Our study covers 48 countries that can be classified into 29 Southern and 19 Northern countries and uses the lead content of gasoline as proxy for environmental stringency. For North-South trade we find significant PH and FE effects going in the expected, opposite directions and being of similar magnitude. However, when looking at world trade, the effects become very small because of the high North-North trade share, where we have no a priori expectations about the signs of these effects. Therefore popular fears about the trade effects of differences in environmental regulations might by exaggerated. The second chapter is entitled "Is trade bad for the Environment? Decomposing worldwide SO2 emissions, 1990-2000". First we construct a novel and large database containing reasonable estimates of SO2 emission intensities per unit labor that vary across countries, periods and manufacturing sectors. Then we use these original data (covering 31 developed and 31 developing countries) to decompose the worldwide SO2 emissions into the three well known dynamic effects (scale, technique and composition effect). We find that the positive scale (+9,5%) and the negative technique (-12.5%) effect are the main driving forces of emission changes. Composition effects between countries and sectors are smaller, both negative and of similar magnitude (-3.5% each). Given that trade matters via the composition effects this means that trade reduces total emissions. We next construct, in a first experiment, a hypothetical world where no trade happens, i.e. each country produces its imports at home and does no longer produce its exports. The difference between the actual and this no-trade world allows us (under the omission of price effects) to compute a static first-order trade effect. The latter now increases total world emissions because it allows, on average, dirty countries to specialize in dirty products. However, this effect is smaller (3.5%) in 2000 than in 1990 (10%), in line with the negative dynamic composition effect identified in the previous exercise. We then propose a second experiment, comparing effective emissions with the maximum or minimum possible level of SO2 emissions. These hypothetical levels of emissions are obtained by reallocating labour accordingly across sectors within each country (under the country-employment and the world industry-production constraints). Using linear programming techniques, we show that emissions are reduced by 90% with respect to the worst case, but that they could still be reduced further by another 80% if emissions were to be minimized. The findings from this chapter go together with those from chapter one in the sense that trade-induced composition effect do not seem to be the main source of pollution, at least in the recent past. Going now to the economic geography part of this thesis, the third chapter, entitled "A Dynamic Model with Sectoral Agglomeration Effects" consists of a short note that derives the theoretical model estimated in the fourth chapter. The derivation is directly based on the multi-regional framework by Ciccone (2002) but extends it in order to include sectoral disaggregation and a temporal dimension. This allows us formally to write present productivity as a function of past productivity and other contemporaneous and past control variables. The fourth chapter entitled "Sectoral Agglomeration Effects in a Panel of European Regions" takes the final equation derived in chapter three to the data. We investigate the empirical link between density and labour productivity based on regional data (245 NUTS-2 regions over the period 1980-2003). Using dynamic panel techniques allows us to control for the possible endogeneity of density and for region specific effects. We find a positive long run elasticity of density with respect to labour productivity of about 13%. When using data at the sectoral level it seems that positive cross-sector and negative own-sector externalities are present in manufacturing while financial services display strong positive own-sector effects. The fifth and last chapter entitled "Is the World's Economic Center of Gravity Already in Asia?" computes the world economic, demographic and geographic center of gravity for 1975-2004 and compares them. Based on data for the largest cities in the world and using the physical concept of center of mass, we find that the world's economic center of gravity is still located in Europe, even though there is a clear shift towards Asia. To sum up, this thesis makes three main contributions. First, it provides new estimates of orders of magnitudes for the role of trade in the globalisation and environment debate. Second, it computes reliable and disaggregated elasticities for the effect of density on labour productivity in European regions. Third, it allows us, in a geometrically rigorous way, to track the path of the world's economic center of gravity.

Examination of Highway Maintenance Garages in the U.S. 30 Corridor between Ames and Cedar Rapids, 1988

A linear programming model is used to optimally assign highway segments to highway maintenance garages using existing facilities. The model is also used to determine possible operational savings or losses associated with four alternatives for expanding, closing and/or relocating some of the garages in a study area. The study area contains 16 highway maintenance garages and 139 highway segments. The study recommends alternative No. 3 (close Tama and Blairstown garages and relocate new garage at Jct. U.S. 30 and Iowa 21) at an annual operational savings of approximately $16,250. These operational savings, however, are only the guidelines for decisionmakers and are subject to the required assumptions of the model used and limitations of the study.

Short-term Hydropower Planning System

The Thesis gives a decision support framework that has significant impact on the economic performance and viability of a hydropower company. The studyaddresses the short-term hydropower planning problem in the Nordic deregulated electricity market. The basics of the Nordic electricity market, trading mechanisms, hydropower system characteristics and production planning are presented in the Thesis. The related modelling theory and optimization methods are covered aswell. The Thesis provides a mixed integer linear programming model applied in asuccessive linearization method for optimal bidding and scheduling decisions inthe hydropower system operation within short-term horizon. A scenario based deterministic approach is exploited for modelling uncertainty in market price and inflow. The Thesis proposes a calibration framework to examine the physical accuracy and economic optimality of the decisions suggested by the model. A calibration example is provided with data from a real hydropower system using a commercial modelling application with the mixed integer linear programming solver CPLEX.

Jatkuvavaletun teräksen leikkaussuunnitelman optimointi

Teollisuuden tuotannon eri prosessien optimointi on hyvin ajankohtainen aihe. Monet ohjausjärjestelmät ovat ajalta, jolloin tietokoneiden laskentateho oli hyvin vaatimaton nykyisiin verrattuna. Työssä esitetään tuotantoprosessi, joka sisältää teräksen leikkaussuunnitelman muodostamisongelman. Valuprosessi on yksi teräksen valmistuksen välivaiheita. Siinä sopivaan laatuun saatettu sula teräs valetaan linjastoon, jossa se jähmettyy ja leikataan aihioiksi. Myöhemmissä vaiheissa teräsaihioista muokataan pienempiä kokonaisuuksia, tehtaan lopputuotteita. Jatkuvavaletut aihiot voidaan leikata tilauskannasta riippuen monella eri tavalla. Tätä varten tarvitaan leikkaussuunnitelma, jonka muodostamiseksi on ratkaistava sekalukuoptimointiongelma. Sekalukuoptimointiongelmat ovat optimoinnin haastavin muoto. Niitä on tutkittu yksinkertaisempiin optimointiongelmiin nähden vähän. Nykyisten tietokoneiden laskentateho on kuitenkin mahdollistanut raskaampien ja monimutkaisempien optimointialgoritmien käytön ja kehittämisen. Työssä on käytetty ja esitetty eräs stokastisen optimoinnin menetelmä, differentiaalievoluutioalgoritmi. Tässä työssä esitetään teräksen leikkausoptimointialgoritmi. Kehitetty optimointimenetelmä toimii dynaamisesti tehdasympäristössä käyttäjien määrittelemien parametrien mukaisesti. Työ on osa Syncron Tech Oy:n Ovako Bar Oy Ab:lle toimittamaa ohjausjärjestelmää.

Identifying strategies for mitigating the global warming impact of the EU-25 economy using a multi-objective input-output approach

Global warming mitigation has recently become a priority worldwide. A large body of literature dealing with energy related problems has focused on reducing greenhouse gases emissions at an engineering scale. In contrast, the minimization of climate change at a wider macroeconomic level has so far received much less attention. We investigate here the issue of how to mitigate global warming by performing changes in an economy. To this end, we make use of a systematic tool that combines three methods: linear programming, environmentally extended input output models, and life cycle assessment principles. The problem of identifying key economic sectors that contribute significantly to global warming is posed in mathematical terms as a bi criteria linear program that seeks to optimize simultaneously the total economic output and the total life cycle CO2 emissions. We have applied this approach to the European Union economy, finding that significant reductions in global warming potential can be attained by regulating specific economic sectors. Our tool is intended to aid policymakers in the design of more effective public policies for achieving the environmental and economic targets sought.

Matemática económica con programación lineal

La percepción del joven estudiante de economía es que la práctica con ejercicios es lo único que debe saber. Ésta percepción se puede cambiar con la Programación Lineal ya que unimos teoría y práctica y, al mismo tiempo, mejoramos la capacidad de modelar situaciones económicas y además, hacemos énfasis en el uso de las matemáticas como herramienta eficaz en la mejora de las actividades propias.