Probabilidad de empate

En relación al empate que hoy se ha producido en la asamblea de la CUP, lamentamos tener que desmentir afirmaciones como esta:



La probabilidad de que se haya producido este empate no es tan baja como indica Mario Bilbao. La probabilidad de que determinado número de asistentes a la asamblea vote por la investidura de Mas es función de la probabilidad p de que cada uno de los miembros de la asamblea desee investir presidente a Mas: los resultados habrían sido muy distintos si todos los asistentes estuvieran convencidos de investir a Mas (p próxima a 1) o de no investirlo (p próxima a 0). Para determinar la probabilidad de que exactamente 1515 de los 3030 asistentes hayan votado por la investidura de Mas, hemos de utilizar la distribución de probabilidad binomial.

El siguiente gráfico muestra la probabilidad de que se produzca empate (prob) en función de la probabilidad de que un miembro de la asamblea desee que Mas sea investido (p):



Como Antoni Ítalo (entre otros) ha indicado, la probabilidad de empate es como máximo del 1,45%, y es (ahora sí) prácticamente despreciable si p no está entre 0,47 y 0,52 aproximadamente.

The Spanish corporate structure through interlocking directorates

Citation

Sicilia, Carlos; Sallan, Jose M.; Simo, Pep (2016). The Spanish corporate structure through interlocking directorates. Cuadernos de Gestión, 63(1):63-84.
DOI: http://dx.doi.org/10.5295/cdg.120336ps

Abstract

The objective of this paper is to determine to what extent globalisation pressures are changing networks modelled on continental capitalism into Anglo-Saxon models. For this purpose we analyse the Spanish network of interlocks, comparing the present structure with that of 1993. We show how, although Spanish corporative structure continues to display characteristics of the continental economies, some major banks are significantly reducing industrial activity. Nevertheless, the financial organisations continue to maintain a close relationship with sectors such as construction and services. According to the theories on the role of the interlocking directorates, the companies in these sectors assure their access to banking credit by incorporating advisors from financial organisations to their board of directors.

El objetivo de este trabajo es determinar en qué medida las presiones de la globalización han cambiado los modelos que siguen las redes corporativas españolas desde los propios del capitalismo continental hacia los modelos anglosajones. Para ello se analiza la red española de consejerías cruzadas, comparando la estructura actual con la de 1993. Se muestra cómo, a pesar de la estructura corporativa española continúa mostrando características de las economías continentales, algunos de los principales bancos están reduciendo significativamente la actividad industrial. No obstante, las entidades financieras siguen manteniendo una estrecha relación con sectores como la construcción y los servicios. De acuerdo con las teorías sobre el papel de las consejerías cruzadas, las empresas de estos sectores tratan de asegurar su acceso al crédito bancario mediante la incorporación de asesores de las entidades financieras a su consejo de administración.

Combinatorial explosion

Combinatorial optimization is about picking one element of a set of finite objects that optimizes an objective function. Put in this way, the solution may seem straightforward: just enumerate all elements of the set and assess the objective function for each object.

The problem with this naïve approach is that the number of elements of the set can be very large. Let's consider, for instance, the symmetric travelling salesperson problem. For an instance of n nodes, the number of possible solutions is the factorial of (n-1) divided by two. This means that, for a problem of size n=20, and taking one millisecond to assess each solution, it takes around 1,108,606 years to assess all solutions. This effect is called combinatorial explosion.

This video shows an example of an even faster combinatorial explosion. Additionally, helps in learning how to count large numbers in Japanese...

Modeling and Solving Linear Programming with R

Citation

Sallan, Jose M.; Lordan, Oriol; Fernandez, Vicenc (2015). Modeling and Solving Linear Programming with R. Omniascience.

Abstract

Linear programming is one of the most extensively used techniques in the toolbox of quantitative methods of optimization. One of the reasons of the popularity of linear programming is that it allows to model a large variety of situations with a simple framework. Furthermore, a linear program is relatively easy to solve. The simplex method allows to solve most linear programs efficiently, and the Karmarkar interior-point method allows a more efficient solving of some kinds of linear programming. The power of linear programming is greatly enhanced when came the opportunity of solving integer and mixed integer linear programming. In these models all or some of the decision variables are integers, respectively.

In this book we provide a brief introduction to linear programming, together with a set of exercises that introduce some applications of linear programming. We will also provide an introduction to solve linear programming in R. For each problem a possible solution through linear programming is introduced, together with the code to solve it in R and its numerical solution.

Availability

Please see Omniascience website for details, and the repository containing the code used in the book.

Also available through UPCommons in this direction.

Correlation does not mean causation II

Just another example of spurious correlation...

An Examination of the Effects of Self-Regulatory Focus on the Perception of the Media Richness: The Case of E-Mail

Citation

Armengol, Xavier; Fernandez, V.; Simo, P.; Sallan, J. M. (2015). An Examination of the Effects of Self-Regulatory Focus on the Perception of the Media Richness: The Case of E-Mail. International Journal of Business Communication
http://dx.doi.org/10.1177/2329488415572780

Abstract

Communication is a key element in organizations’ business success. The media richness theory and the channel expansion theory are two of the most influential theories regarding the selection and use of communication media in organizations; however, literature has focused little on the effects of self-regulation by managers and employees in these theories. To analyze these topics, this study develops an empirical investigation by gathering data from 600 managers and employees using a questionnaire. The results suggest that the perception of media richness is positively affected when the individual shows a promotion focus or strategy.

Correlation does not mean causation

A variety of images showing that correlation does not imply causation.

A plot showing the relationship between the number of Nobel prizes earned in each country, and chocolate consumption. Both data are adjusted by the population of each country (an indirect source of this graph here).

Phenomena like the shown in the image above usually come from the fact that both variables involved correlate highly with a third (not shown) variable. Many times this third variable is population (source: xkcd)

The fact that a linear association might be easily confounded with a causal relationship may be the reason of the mith that storks deliver babies. This article shows the strong correlation between stork population and human birth rate.

Conclusion, keep in mind that correlation does not mean causation. Although an unexpected correlation may lead to a serendipious discovering (source: xkcd)...

50 shades of grey

The readers of the novel series on unconventional desires may enjoy this little picture showing (literally) 50 shades of grey:

The picture is easily generated using the following R code:

shades <- seq(0, 1, length.out =50)
m.shades <- matrix(rep(shades, 50), 50, 50)
shades.of.grey <- grey(shades)
png("shades.png")
image(m.shades, col=shades.of.grey)
dev.off()

Estudio de un plan de operaciones para una nueva modalidad de handling en la terminal T1 del aeropuerto BCN-El Prat

Student

Àlvar Jaime Hileno

Supervisors

Oriol Lordan / Jose M Sallan

Defense date 

30th June, 2014

Abstract

Carrying the luggage to the airport on your way back home can be an exhausting and stressing task, and some customers of luxury hotels may be interested in a service in which their luggage is dropped directly in the hotel.

This project assesses the operational, legal and financial viability of a business consisting in facilitating the check in directly in the hotel, and carrying customer's luggage directly to the airport handling services in the context of the El-Prat Barcelona (BCN) airport. 

Estudio de implantación de un Pace-Maker en una línea de montaje de una empresa del sector de la automoción

Student

Mireia Magallón Gracia

Supervisors

Jose M Sallan

Defense date 

12th February, 2015

Abstract

English title: Study of the implementation of a pacemaker in an assembly line in the automotive industry

One of the tools that can be used in the context of lean manufacturing is a pacemaker. In this project several alternatives to implement a pacemaker in an assembly line are assessed. Then, the guidelines for implementing the selected alternative are detailed, and the implementation process is planned through the PMBOK standards in project management.

A list of linear programming solvers

Here is a list of software available to solve linear programming (LP) models, focused on free software with an R implementation. Most of this software is also able to solve integer linear programming (ILP) and mixed integer linear programming (MILP) models. The list is intended to be extended and more extensively annotated in the future. Don't hesitate to comment below if you see how this list can be extended or enhanced.

lp_solve

lp_solve is a MILP solver, that runs in a large variety of languages. Models can be passed via input files, an API (application programming interface) or an IDE (integrated development environment). It is also embedded into R via the lp_solve and lp_solveAPI packages. The reference guide is quite exhaustive, and a good place to start using this software.

lp_solve reference guide
lp_solve FAQ (MIT)
R package lp_solve
R package lp_solveAPI

GLPK

GLPK stands for GNU Linear Programming Kit. It is a set of callable libraries written in C intended to solve large scale LP, ILP and MILP models. It is developed by Andrei Makhorin, of the Moscow Aviation Institute. GLPK contains an standalone solver, glpsol, that can be called from the command line. It can read models written in a variety of languages, among them CPLEX, convenient for relatively small problems and GNU MathProg, a standard based on the AMPL format adequate to write large models with regular structure.

GLPK can be used in R via the Rglpk package. For windows users is available GUSEK, an IDE running the GLPK libraries.

GLPK homepage
R package Rglpk
GUSEK IDE

SYMPHONY

Symphony is an open-source solver of MILP problems written in C. It is an initiative of the COIN-OR project. It can read problems written in MPS and MathProg formats. It is also implemented in R through the Rsymphony package.

SYMPHONY homepage
SYMPHONY user's manual
R package Rsymphony

linprog

A library for solving linear programming models, quite popular in MATLAB. There is also a R package available. This library does not solve ILP/MILP models.

linprog in MATLAB
linprog R package

Análisis longitudinal de las consejerías cruzadas y su papel en la organización económica española

Citation

Poveda, G.; Sicilia, C.; Simo, P.; Sallan, J. M. (2015). Análisis longitudinal de las consejerías cruzadas y su papel en la organización económica española. Harvard-Deusto Business Research, 3(2):64-73.
http://dx.doi.org/10.3926/hdbr.7

Abstract

Interlocking directorates build relationships between corporations, and define the national economic organization. We analyze the corporate network of the IBEX 35 companies, plus two savings banks, and compare it with the same network in 2005. The most central positions in the network are occupied by large Spanish corporations, privatized in the 90s. On the other hand, in 2008 banks lose centrality, while savings bank’s centrality rises.