Search

El propósito de esta investigación es desarrollar y validar un sistema inteligente para la predicción eficaz de la disponibilidad de plazas de estacionamiento en contextos urbanos, específicamente en estacionamientos, enfocándose en la integración de técnicas avanzadas de aprendizaje profundo y análisis temporal de series de datos.

Membrane computing is a massively parallel and non-deterministic bioinspired computing paradigm whose models are called P systems. Validating and testing such models is a challenge which is being overcome by developing simulators. Regardless of their heterogeneity, such simulators require to read and interpret the models to be simulated. To this end, P-Lingua is a high-level P system definition language which has been widely used in the last decade. The P-Lingua ecosystem includes not only the language, but also libraries and software tools for parsing and simulating membrane computing models. Each version of P-Lingua supported new types or variants of P systems. This leads to a shortcoming: Only a predefined list of variants can be used, thus making it difficult for researchers to study custom ones. Moreover, derivation modes cannot be user-defined, i.e, the way in which P system computations should be generated is determined by the simulation algorithm in the source code. The main contribution of this paper is a completely new design of the P-Lingua language, called P-Lingua 5, in which the user can define custom variants and derivation modes, among other improvements such as including procedural programming and simulation directives. It is worth mentioning that it has backward-compatibility with previous versions of the language. A completely new set of command-line tools is provided for parsing and simulating P-Lingua 5 files. Finally, several examples are included in this paper covering the most common P system types.

Membrane computing is an unconventional computing paradigm that has gained much attention in recent decades because of its massively parallel character and its usefulness to build models of complex systems. However, until now, there was no generic hardware implementation of P systems. Computational frameworks to execute P systems up to this day rely on the simulation of the parallel working mechanisms of P systems by inherently sequential algorithms. Such algorithms can then be implemented as is or can be parallelized, up to a certain point, to run on parallel computers. However, this is not as efficient as a dedicated parallel hardware implementation. There have been ad hoc implementations of particular P systems for parallel hardware, but they lack to be problem-generic or they are not scalable enough to implement large P systems. In this paper, a first intrinsically parallel hardware architecture to implement generic P system models is introduced. It is designed to be straightforwardly implemented in programmable logic circuits like FPGAs. The feasibility and correct execution of our architecture has been verified by means of a simulator, and several simulation results for different P system examples have been analysed to foresee the pros and cons of this design.

Este trabajo se basa en el desarrollo de una herramienta de detección de plagio en código R y RMarkdownn. Con el fin de ofrecer al personal docente un software de código abierto y On Premise capaz de detectar plagio en entregas con esta extensión. Con la finalidad de ofrecer un software de código abierto capaz de detectar plagio entre documentos con esta codificación, brindando la posibilidad al personal docente de detectar plagio en entregas con extensión .R y .Rmd. El contenido de este trabajo se ha dividido en 7 capítulos de la siguiente forma: En el capítulo 1 se realiza una introducción sobre el plagio, el plagio en código fuente y como se combate el plagio dentro de las universidades. A lo largo del capítulo 2 se describen los aspectos generales del estado del arte, se analizan diferentes programas de detección de plagio en código fuente, destacando sus principales cualidades y analizando sus carencias. Finalmente, se realiza un análisis de dicho estado del arte, sirviendo como argumentación fundamental para el desarrollo de la herramienta de software que se presenta en este TFG. El capítulo 3 se centra en el paquete SimilaR, ya presentado en el capítulo anterior. Este paquete es quizás el más utilizado actualmente para la detección de plagio en código R. En este capítulo se estudian minuciosamente sus prestaciones y limitaciones, con la finalidad de proponer mejoras. El capítulo 4 está dedicado a explicar los detalles de la nueva implementación. Se consigue que la nueva herramienta, que bautizamos como SimilaR2, sea capaz de detectar plagio en secciones de comentarios gracias a la implementación de algoritmos fingerprints, así como ampliar las zonas de código R en dónde se puede detectar plagio (SimilaR tan solo puede detectar plagio en el cuerpo de funciones, pero no en el código escrito fuera de ellas). Adicionalmente, se explica la salida de SimilaR2, la cual aporta información detallada del análisis de detección de plagio, análisis de clústering y nubes d, This eassay based on the development of a plagiarism detection tool in R code and RMarkdownn. The aim is to provide teaching staff with an open source and On Premise software capable of detecting plagiarism in deliveries with this extension. The content of this work has been divided into 7 chapters as follows: Chapter 1 provides an introduction to plagiarism, plagiarism in source code and how plagiarism is combated within universities. Chapter 2 describes the general aspects of the state of the art analyses different programmes for detecting plagiarism in source code, highlighting their main qualities and analysing their shortcomings. Finally, an analysis of the state of the art is carried out, serving as a fundamental argumentation for the development of the software tool presented in this dissertation. Chapter 3 focuses on the SimilaR package, already presented in the previous chapter. This package is perhaps the most widely used at present for the detection of plagiarism in R code. In this chapter, its features and limitations are studied in detail, with the aim of proposing improvements. Chapter 4 is dedicated to explain the details of the new implementation. The new tool, which we are going to call SimilaR2, is able to detect plagiarism in comment sections thanks to the implementation of fingerprinting algorithms, as well as to extend the areas of R code where plagiarism can be detected (SimilaR can only detect plagiarism in the body of functions, but not in the code written outside them). Additionally, the output of SimilaR2 is explained, which provides detailed information on plagiarism detection analysis, clustering and word cloud analysis, among other possibilities. During chapter 5, the process of building the SimilaR2 R package is detailed, as well as an interactive web application developed with Shiny, the latter being able to run SimilaR2 and display the results in a web interface. In chapter 6, a statistical analysis of the results obtained by the Simi

To date, parallel simulation algorithms for spiking neural P (SNP) systems are based on a matrix representation. This way, the simulation is implemented with linear algebra operations, which can be easily parallelized on high performance computing platforms such as GPUs. Although it has been convenient for the first generation of GPU-based simulators, such as CuSNP, there are some bottlenecks to sort out. For example, the proposed matrix representations of SNP systems lead to very sparse matrices, where the majority of values are zero. It is known that sparse matrices can compromise the performance of algorithms since they involve a waste of memory and time. This problem has been extensively studied in the literature of parallel computing. In this paper, we analyze some of these ideas and apply them to represent some variants of SNP systems. We also provide a new simulation algorithm based on a novel compressed representation for sparse matrices. We also conclude which SNP system variant better suits our new compressed matrix representation.

Nature-inspired computing is a type of human-designed computing motivated by nature, which is based on the employ of paradigms, mechanisms, and principles underlying natural systems. In this article, a versatile and vigorous bio-inspired branch of natural computing, named membrane computing is discussed. This computing paradigm is aroused by the internal membrane function and the structure of biological cells. We first introduce some basic concepts and formalisms of membrane computing, and then some basic types or variants of P systems (also named membrane systems) are presented. The state-of-the-art computability theory and a pioneering computational complexity theory are presented with P system frameworks and numerous solutions to hard computational problems (especially NP-complete problems) via P systems with membrane division are reported. Finally, a number of applications and open problems of P systems are briefly described.

Dendrite P systems (DeP systems) are a recently introduced neural-like model of computation. They provide an alternative to the more classical spiking neural (SN) P systems. In this paper, we present the first software simulator for DeP systems, and we investigate the key features of the representation of the syntax and semantics of such systems. First, the conceptual design of a simulation algorithm is discussed. This is helpful in order to shade a light on the differences with simulators for SN P systems, and also to identify potential parallelizable parts. Second, a novel simulator implemented within the PLingua simulation framework is presented. Moreover, MeCoSim, a GUI tool for abstract representation of problems based on P system models has been extended to support this model. An experimental validation of this simulator is also covered.

A mobile robot acting in a human environment should follow social conventions, keeping safety distances and navigating at moderate speeds, in order to respect people in its surroundings and avoid obstacles in real-time. The problem is more complex in differential-drive wheeled robots, with trajectories constrained by nonholonomic and kinematics restrictions. It is an NP-hard problem widely studied in the literature, combining disciplines such as Psychology, Mathematics, Computer Science and Engineering. In this work, we propose a novel solution based on Membrane Computing, Social Force Model and Dynamic Window Approach Algorithm. The resulting model is able to compute, in logarithmic time, the best motion command for the robot, given its current state, considering the surrounding people and obstacles. The model is compatible with other membrane computing models for robotics and suitable for an implementation on parallel hardware. Finally, a visual simulator was implemented in ROS and C++ for validation and testing.

Psystems with active membranes is one of the most studied models within the field ofMembrane Computing. Simulating the organization and behavior of the living cells througha tree-like structure and abstracting the mechanisms that help the cell to keep alive intorules (evolution, communication, dissolution and division rules), they have been used tosolve several computationally hard problems. We are dealing with non-cooperative systemshere, that is, the number of reactives in a rule is always one. Even then, it has been proventhat problems from the classPSPACEcan be solved, so in order to acquire a minimal modelthat can solve computationally hard problems, polarizations are removed. In this paper wefind the relevance of the length of the right-hand side of the rule, being necessary whenusing separation rules and being irrelevant when division rules are used, improving somesolutions previously presented, restricting the right-hand side of the rules, obtaining newfrontiers of efficiency in this framework. The state of the art of these systems is presentedin a graphical way.

In recent years, incremental sampling-based motion planning algorithms have been widely used to solve robot motion planning problems in high-dimensional configuration spaces. In particular, the Rapidly-exploring Random Tree (RRT) algorithm and its asymptotically-optimal counterpart called RRT* are popular algorithms used in real-life applications due to its desirable properties. Such algorithms are inherently iterative, but certain modules such as the collision-checking procedure can be parallelized providing significant speedup with respect to sequential implementations. In this paper, the RRT and RRT* algorithms have been adapted to a bioinspired computational framework called Membrane Computing whose models of computation, a.k.a. P systems, run in a non-deterministic and massively parallel way. A large number of robotic applications are currently using a variant of P systems called Enzymatic Numerical P systems (ENPS) for reactive controlling, but there is a lack of solutions for motion planning in the framework. The novel models in this work have been designed using the ENPS framework. In order to test and validate the ENPS models for RRT and RRT*, we present two ad-hoc implementations able to emulate the computation of the models using OpenMP and CUDA. Finally, we show the speedup of our solutions with respect to sequential baseline implementations. The results show a speedup up to 6x using OpenMP with 8 cores against the sequential implementation and up to 24x using CUDA against the best multi-threading configuration.

Motivated by the mechanism of inhibitory synapses, a new kind of spiking neural P (SNP) system rules, called inhibitory rules, is introduced in this paper. Based on this, a new variant of SNP systems is proposed, called spiking neural P systems with inhibitory rules (SNP-IR systems). Different from the usual firing rules in SNP systems, the firing condition of an inhibitory rule not only depends on the state of the neuron associated with the rule but also is related to the states of other neurons. Moreover, from the perspective of topological structure, the new variant is shown as a directed graph with inhibitory arcs, and therefore seems to have more powerful control. The computational completeness of SNPIR systems is discussed. In particular, it is proved that SNP-IR systems are Turing universal number accepting/generating devices. Moreover, we obtain a small universal function-computing device for SNP-IR systems consisting of 100 neurons.

P system simulators are critical tools to enable them as formal modeling framework for real-life applications. Such simulators abstract the concept of P systems in various ways, depending on the needs of the users and the requirements of the specific application. We identify three main levels of abstraction: graphical user interfaces, simulation engines and parallel imple-mentations. In this paper, we survey the state of the art at these levels and discuss the main challenges under consideration for future developments.

Presumably efficient computing models are characterized by their capability to provide polynomial-time solutions for NPcomplete problems. Given a classRof recognizer membrane systems,Rdenotes the set of decision problems solvable by families from R in polynomial time and in a uniform way. PMCR is closed under complement and under polynomial-time reduction. +erefore, if R is a presumably efficient computing model of recognizer membrane systems, then NP ∪ co-NP ⊆PMCR. In this paper, the lower bound NP ∪ co-NP for the time complexity class PMCR is improved for any presumably efficient computing model R of recognizer membrane systems verifying some simple requirements. Specifically, it is shown that DP ∪ co-DP is a lower bound for such PMCR, where DP is the class of differences of any two languages in NP. Since NP ∪ co-NP⊆DP ∩ co- DP, this lower bound for PMCR delimits a thinner frontier than that with NP ∪ co-NP.

In the Membrane Computing area, P systems are unconventional devices of computation inspired by the structure and processes taking place in living cells. Main successful P system applications lie in computability and computational complexity theories, as well as in biological modelling. Given that models become too complex to deal with, simulators for P systems are essential tools and their efficiency is critical. In order to handle the diverse situations that may arise during the computation, these simulators have to take into account that worst-case scenarios can happen, even though they rarely occur. As a result, there is a significant loss of performance. In this paper, the concept of adaptative simulation for P systems is introduced to palliate this problem. This is achieved by passing high-level information provided directly by P system model designers to the simulator, helping it to better adapt to the target model. For this purpose, an existing simulator for an ecosystem modelling framework, named Population Dynamics P systems, is extended to include the information of modules, that are usually employed to define ecosystem models. Moreover, the standard description language for P systems, P-Lingua, has been re-engineered in its version 5. It now includes a new syntactical item, called feature, to express this kind of high-level semantic information. Experiments show that this simple adaptative simulator supporting modules as features doubles the performance when running on GPUs and on multicore processors.

Population Dynamics P systems is a modelling framework that have been used successfully for some important real ecosystems. This model is inherently probabilistic, and the scheme of rules is very exible, allowing even cooperation between membranes. Thus, its simulation has been a challenge in the past years, leading to several simulation algorithms. The latest one, which has been proved to be the most accurate so far, is DCBA. The main drawback of DCBA is its complexity, requiring a very large table to handle all competitions. In this paper, we discuss two strategies to decrease this table, allowing a more lightweight version of DCBA that can be used in parallel implementations.

First, the dependency graph technique, not so far from its current application, was developed trying to nd the shortest computations for membrane systems solving instances of SAT. Certain families of membrane systems have been demonstrated to be non-effcient by means of the reduction of nding an accepting computation (respectively, rejecting computation) to the problem of reaching from a node of the dependency graph to another one. In this paper, a novel application to this technique is explained. Supposing that a problem can be solved by means of a kind of membrane systems leads to a contradiction by means of using the dependency graph as a reasoning method. In this case, it is demonstrated that a single system without dissolution, polarizations and cooperation cannot distinguish a single object from more than one object. An extended version of this work will be presented in the 20th International Conference on Membrane Computing.

P-Lingua is a software framework for Membrane Computing, it includes a programming language, also called P-Lingua, for writting P system de nitions using a syntax close to standard scienti c notation. The rst line of a P-Lingua le is an unique identi er de ning the variant or model of P system to be used, i.e, the semantics of the P system. Software tools based on P-Lingua use this identi er to select a simulation algorithm implementing the corresponding derivation mode. Derivation modes de ne how to obtain a con guration Ct+1 from a con guration Ct. This information is usually hard-coded in the simulation algorithm. The P system model also de nes what types or rules can be used, the P-Lingua compiler uses the identi er to select an speci c parser for the le. In this case, a set of parsers is codi ed within the compiler tool. One for each unique identi er. P-Lingua has grown during the last 12 years, including more and more P system models. From a software engineering point of view, this approximation implies a continous development of the framework, leading to a monolithic software which is hard to debug and maintain. In this paper, we propose a new software approximation for the framework, including a new syntax for de ning rule patterns and derivation modes. The P-Lingua users can now de ne custom P system models instead of hard-coding them in the software. This approximation leads to a more exible solution which is easier to maintain and debug. Moreover, users could de ne and play with new/experimental P system models.

In 2008, P-Lingua was born. The Research Group on Natural Computing worked on the development of simulation tools since the beginning of Membrane Computing. However, back in 2007, researchers from the group set out to the ambitious journey of creating a generic simulation framework for P systems. P-Lingua has evolved since then, o ering more exibility and a wider range of supported models. Many applications have also branched from this software project. In this paper, we brie y survey the evolution of P-Lingua to date, some of the associated applications, and prospective paths for upcoming challenges in the research area.

The dependency graph technique was initially thought as a method to find short paths in the computation tree of a membrane system using weak metrics. It could be used to obtain reasonably fast SAT-solvers, capable of competing with the ones available in the literature. Later on, they were used as a method to demonstrate the non-efficiency of some membrane systems, capturing the dynamics of the systems by a static directed graph structure. Recently, the dependency graphs have also been used to establish negative results in Membrane Computing. Specifically, in this work, demonstrating the inability of a kind of membrane system to solve some decision problems efficiently by means of a single system.

Membrane computing is a bio-inspired computing paradigm that lacks in vivo implementation. That is why software or hardware implementations have to be used to validate models. Several tools have been created for this purpose; some of them are created for specific purposes, such as solving a computationally hard problem; and others are more generic, to cover a broad spectrum of possible models. The former have the advantage of being very efficient, crucial for solving large instances of certain problems; however, this efficiency leads to a loss of generality, since algorithms are usually hard-coded and they do not allow other models. On the contrary, the latter are perfect tools for researchers, given that new models can be checked without much effort by defining them in the framework; since these algorithms have to simulate as many models as possible, they lack specificities to improve the performance. P-Lingua has been widely used to simulate membrane systems, having integrated both a language and a simulator. To obtain better results in terms of time used to simulate models defined in this language, a new perspective is studied. The model defined in P-Lingua will be compiled into C++ source code that will implement an ad hoc simulator. This code will consider specifications about how rules have to be executed, that is, some simple specifications of the semantics. To show how it works, some examples of specifications of models will be presented, which can be simulated using the new-developed GNU GPLv3 command-line tool pcc.

It is a long-cherished wish to implement numerical P systems (NPS) on a parallel architecture so that its large scale parallelism can be exploited to speedup computation tremendously. FPGA is a reconfigurable hardware in which operations are triggered so synchronized by edge or level of activating signals, making it an eligible platform to implement NPS and its variant, enzymatic numerical P system (ENPS). In this article, a NPS and a ENPS designed as robot controllers are implemented in FPGA, achieving a speedup of 105 comparing to software simulation. FPGA hardened NPS in this research can be regarded as a heterogeneous multicore processor since membranes inside work as processing units which possess different functions. FPGA hardened NPS is imparted universal asynchronous receiver/transmitter (UART) communication ability to push it closer to real-life application. FPGA hardened ENPS consume less hardware resources and power for less complicate membrane structures and processes.

Solving the feasibility problem in robotic motion planning means to find feasible trajectories for specific mobile robots acting in environments with obstacles whose positions are known a priori. The Rapidly-exploring Random Tree (RRT) algorithm is a classical algorithm to solve such a problem in real-life applications. In this paper, we provide a model in the framework of Enzymatic Numerical P systems to reproduce the behaviour of the RRT algorithm. A C++ ad-hoc simulator is also provided to validate the model.

A short summary of recent ecological modelling applications within mem- brane computing developed by the Research Group on Natural Computing (RGNC) is presented here. A timeline of models presented in the last 10 years is provided, together with some comments on the associated software tools

In order to validate P system models and to assist on their formal verification, simulators are indispensable. Moreover, having effi-cient simulation tools is crucial, and for this purpose, parallel platforms should be employed. So far, several parallel simulators for P systems have been developed, specifically targeting GPUs (Graphics Processing Units). Although being a hot topic within Membrane Computing, map-ping P system parallelism on GPUs is still not a mature area. In the past, we have successfully accelerated the simulation of two specific fam-ilies of P systems solving SAT with GPUs, and learned in the process some semantics ingredients that fit well on these parallel devices. We are extending this exploration by designing an specific simulator of a P system model for the FACTORIZATION problem. In this paper, we analyse the two main approaches for simulators, and depict some design decisions required for this case study.

P system simulators are of high importance in Membrane Computing, since they provide tools to assist on model validation and verification. Keeping a balance between generality and flexibility, on the one side, and efficiency, on the other hand, is always challenging, but it is worth the effort. Besides, in order to prove the feasibility of P system models as practical tools for solving problems and aid in decision making, it is essential to provide functional mechanisms to have all the elements required at disposal of the potential users smoothly integrated in a robust workflow. The aim of this paper is to describe the main components and connections within the approach followed in this pipeline.

Methods based on Rapidly-exploring Random Trees (RRTs) have been widely used in robotics to solve motion planning problems. On the other hand, in the membrane computing framework, models based on Enzymatic Numerical P systems (ENPS) have been applied to robot controllers, but today there is a lack of planning algorithms based on membrane computing for robotics. With this motivation, we provide a variant of ENPS called Random Enzymatic Numerical P systems with Proteins and Shared Memory (RENPSM) addressed to implement RRT algorithms and we illustrate it by simulating the bidirectional RRT algorithm. This paper is an extension of [21]a. The software presented in [21] was an ad-hoc simulator, i.e, a tool for simulating computations of one and only one model that has been hard-coded. The main contribution of this paper with respect to [21] is the introduction of a novel solution for membrane computing simulators based on automatic programming. First, we have extended the P-Lingua syntax –a language to define membrane computing models– to write RENPSM models. Second, we have implemented a new parser based on Flex and Bison to read RENPSM models and produce source code in C language for multicore processors with OpenMP. Finally, additional experiments are presented.

Methods based on Rapidly-exploring Random Trees (RRTs) have been in use in robotics to solve motion planning problems for nearly two decades. On the other hand, in the membrane computing framework, models based on Enzymatic Numerical P systems (ENPS) have been applied to robot controllers. These controllers handle the power of motors according to motion commands usually generated by planning algorithms, but today there is a lack of planning algorithms based on membrane computing for robotics. With this motivation, we provide a new variant of ENPS called Random Enzymatic Numerical P systems with Proteins and Shared Memory (RENPSM) addressed to implement RRT algorithms and we illustrate it by presenting a model for path planning of mobile robots based on the bidirectional RRT algorithm. A software for RENPSM has been developed within the Robot Operating System (ROS) and simulation experiments have been conducted by means of the Pioneer 3-DX robot simulation platform.

The P versus NP problem is undoubtedly the most important open question in computer science. Frontiers of tractability or efficiency expressed by means of syntactic or semantic ingredients in the framework of Membrane Computing, an unconventional computing paradigm, can bring a new approach to tackle P versus NP. In this context, the role of the cooperation of objects to trigger rewriting rules is analysed in order to obtain this kind of borderlines. Besides, a relationship among cooperative rewriting rules and instances of 2–SAT problem and 3–SAT problem is highlighted and their connections with results of computational complexity theory are described

Methods based on Rapidly-exploring Random Trees (RRTs) have been in use in robotics to solve motion planning problems for nearly two decades. On the other hand, models based on Enzymatic Numerical P systems (ENPS) have been applied to robot controllers for more than six years. These controllers in real robots handle the power of motors ac- cording to motion commands usually generated by planning algorithms, but today there is a lack of planning algorithms based on membrane sys- tems for robotics. With this motivation, we provide in this paper a new variant of ENPS called Random Enzymatic Numerical P systems with Proteins and Shared Memory (RENPSM) oriented to RRTs for planning in robotics and we illustrate it by presenting a model for generation of RRTs with holonomic limitations. We are working on the ENPS frame- work with the idea of moving towards a complete mobile robot system based on membrane systems, i.e. including controllers and planning; and we have incorporated new ingredients into the ENPS framework to meet the requirements of the RRT generation algorithm.