DOCSLIB.ORG

Evolutionary Computation Swarm Intelligence and Particle Swarm Optimisation

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text

Load more

Evolutionary Computation Evolutionary Computation Swarm Intelligence and Particle Swarm Optimisation Shan He School for Computational Science University of Birmingham Modules 02411 and 22313: EC Evolutionary Computation Outline Outline of Topics Swarm behaviour BOID model Particle Swarm Optimiser Evolutionary Computation Swarm behaviour What is Swarm in nature? Picture from NY Times Evolutionary Computation Swarm behaviour What is Swarm in nature? I Starlings swarm in Rome I Sardine school Evolutionary Computation Swarm behaviour What is Swarm Intelligence? I General definition: Collective behaviour of decentralized, self-organized natural or artificial systems I Self-organized system: an initially disordered system, but some form of global order or coordination will arise out of the local interactions between its components I A swarm intelligence system typically consists of simple agents interacting with each other and their environment following simple rules I Emergence: global order or coordination emerges from these simple interacting rules Evolutionary Computation Swarm behaviour Why swarm? I There is no clear, unique explanation to the evolution of swarm behaviour. I Some possible explanations: I Better defence, e.g., early warning, confuse predator I Better foraging efficiency I Saving energy, e.g., wile geese fly in formation (wedge or skein) I Outcomes of natural selection Evolutionary Computation Swarm behaviour What is Swarm Intelligence in CS? I In CS, we define Swarm Intelligence as \a multi-agent system that has self-organized behaviour that shows some intelligent behaviour." I Two main research fields: I Swarm Intelligence algorithms: a variety of novel algorithms inspired by natural swarm system: I Ant colony optimisation: a SI algorithm inspired by ants' behaviour, useful to solve combinatorial optimisation problems, such as finding better paths through graphs. I Particle Swarm Optimisation I Swarm Robotics: design coordination of multirobot systems which consist of large numbers of mostly simple physical robots. I Goes beyond distributed system, aims to promote scalability, e.g., controlling a large number of robots I http://www.swarmanoid.org/ Evolutionary Computation BOID model BOID I Invented In 1986 by Craig Reynolds to simulate coordinated animal motion such as bird flocks and fish schools. I Winner of technical Oscar: 1997 Sci-Tech Awards from Academy of Motion Pictures and Science I Each individual of the model is called boid which is manoeuvred by three simple predefined behaviours: Evolutionary Computation BOID model Separation I Separation: Steer to avoid crowding local neighbour Evolutionary Computation BOID model Alignment I Alignment: steer towards the average heading of local neighbour Evolutionary Computation BOID model Cohesion I Cohesion: steer to move toward the average position of local neighbour Evolutionary Computation BOID model Demonstration I BOID models have been used in many films I One famous example Batman Returns (1992): BOID was used to create bat swarms and "army" of penguins marching through the streets of Gotham City I Amazing BOID model I Made with Cinder I If you want to learn Cinder to create your own model: Tutorial from Robert Hodgin Evolutionary Computation Particle Swarm Optimiser Particle Swarm Optimiser (PSO) I Invented by Kennedy and Eberhart 1995 I Inspired by bird flocking and fish schooling, more precisely, BIOD I Simple rules for searching global optima I Primarily for real-valued optimisation problems I Simpler but sometimes better than GAs Evolutionary Computation Particle Swarm Optimiser PSO: detailed algorithm I Can be seen as a swarm of particles flying in the search space to find the optimal solution. I Each particle is a solution to the problem, represented by t position Xi in the search space t+1 I Each particle also has a velocity Vi , used to update its position I The variation operator consists of only two equations: t+1 t t t Vi = !Vi + c1r1(Pi − Xi ) + c2r1(Pg − Xi ) (1) t+1 t t+1 Xi = Xi + Vi (2) Pi is the best previous position of the ith particle; Pg is the global best position of the swarm; ! 2 (0; 1] is inertia weigh; c1 and c2 are constants, or so-called learning factors; r1 and r2 are random number in the range of (0; 1) Evolutionary Computation Particle Swarm Optimiser Generic PSO Algorithm X0 := generate an initial population of M particles (solutions) P0 := X0 terminationflag := false t := 0 Evaluate the fitness of each particle in X0. while (terminationflag != true) t Select the best particle Pg from X based on their fitness. For i = 1 : M t t If Xi is better than Pi then Pi := Xi ; // Update Pi t+1 Calculate Vi according equation (1) t+1 Update Xi according equation (2) End For Evaluate the fitness of each particle in Xt+1. t := t + 1 If a termination criterion is met: terminationflag := true Output Pg Evolutionary Computation Particle Swarm Optimiser PSO: algorithm illustration I The search direction of PSO is determined by: I The autobiographical memory, which remembers the best previous position of each individual Pi in the swarm I The publicized knowledge, which is the best solution Pg currently found by the population Evolutionary Computation Particle Swarm Optimiser PSO: practical issues I One key problem faced by EC researcher is how to choose appropriate operators and parameters I PSO: Plug-and-play optimisation algorithm I Only 3 parameters, all not very sensitive I ! can be a linearly decreasing value - better local search I Download my source code at here, you can visualise how PSO find optima..
Recommended publications
  • Evolving Behaviour Trees for Supervisory Control of Robot Swarms

    Evolving Behaviour Trees for Supervisory Control of Robot Swarms

    Evolving Behaviour Trees for Supervisory Control of Robot Swarms Elliott Hogg1y, Sabine Hauert1, David Harvey, Arthur Richards1 1Bristol Robotics Laboratory, University of Bristol, UK. E-mail: [email protected], [email protected] Abstract: Supervisory control of swarms is essential to their deployment in real world scenarios to both monitor their operation and provide guidance. We explore mechanisms by which humans might provide supervisory control to swarms and improve their performance. Rather than have humans guess the correct form of supervisory control, we use artificial evolution to learn effective human-readable strategies to help human operators understanding and performance when con- trolling swarms. These strategies can be thoroughly tested and can provide knowledge to be used in the future in a variety of scenarios. Behaviour trees are applied to represent human readable decision strategies which are produced through evolution. We investigate a simulated set of scenarios where a swarm of robots have to explore varying environments and reach sets of objectives. Effective supervisory control strategies are evolved to explore each environment using different local swarm behaviours. The evolved behaviour trees are animated alongside swarm simulations to enable clear under- standing of the supervisory strategies. We conclude by identifying the strengths in accelerated testing and the benefits of this approach for scenario exploration and training of human operators. Keywords: Human Swarm Interaction; Swarm; Artificial Evolution; Supervisory Control; Behaviour Trees 1. INTRODUCTION • Parametric: Adjusting the characteristics of an emer- gent behaviour [6, 12]. Growing interest in the use of swarm systems has led • Environmental: The operator highlights points of inter- to new questions regarding effective design for real-world est within an environment [8, 26].
  • Mobility and Infectiousness in the Spatial Spread of an Emerging Fungal Pathogen

    Mobility and Infectiousness in the Spatial Spread of an Emerging Fungal Pathogen

    Received: 7 May 2020 | Accepted: 11 January 2021 DOI: 10.1111/1365-2656.13439 RESEARCH ARTICLE Mobility and infectiousness in the spatial spread of an emerging fungal pathogen Kate E. Langwig1 | J. Paul White2 | Katy L. Parise3 | Heather M. Kaarakka2 | Jennifer A. Redell2 | John E. DePue4 | William H. Scullon5 | Jeffrey T. Foster3 | A. Marm Kilpatrick6 | Joseph R. Hoyt1 1Department of Biological Sciences, Virginia Polytechnic Institute, Blacksburg, VA, USA; 2Wisconsin Department of Natural Resources, Madison, WI, USA; 3Pathogen and Microbiome Institute, Northern Arizona University, Flagstaff, AZ, USA; 4Michigan Department of Natural Resources, Baraga, MI, USA; 5Michigan Department of Natural Resources, Norway, MI, USA and 6Department of Ecology and Evolutionary Biology, University of California, Santa Cruz, CA, USA Correspondence Kate E. Langwig Abstract Email: [email protected] 1. Emerging infectious diseases can have devastating effects on host communities, caus- Funding information ing population collapse and species extinctions. The timing of novel pathogen arrival National Science Foundation, Grant/Award into naïve species communities can have consequential effects that shape the trajec- Number: DEB- 1115895, DEB- 1336290 and DEB- 1911853; U.S. Fish and Wildlife tory of epidemics through populations. Pathogen introductions are often presumed Service, Grant/Award Number: F17AP00591 to occur when hosts are highly mobile. However, spread patterns can be influenced Handling Editor: Isabella Cattadori by a multitude of other factors including host body condition and infectiousness. 2. White- nose syndrome (WNS) is a seasonal emerging infectious disease of bats, which is caused by the fungal pathogen Pseudogymnoascus destructans. Within- site transmission of P. destructans primarily occurs over winter; however, the influence of bat mobility and infectiousness on the seasonal timing of pathogen spread to new populations is unknown.
  • Expert Assessment of Stigmergy: a Report for the Department of National Defence

    Expert Assessment of Stigmergy: a Report for the Department of National Defence

    Expert Assessment of Stigmergy: A Report for the Department of National Defence Contract No. W7714-040899/003/SV File No. 011 sv.W7714-040899 Client Reference No.: W7714-4-0899 Requisition No. W7714-040899 Contact Info. Tony White Associate Professor School of Computer Science Room 5302 Herzberg Building Carleton University 1125 Colonel By Drive Ottawa, Ontario K1S 5B6 (Office) 613-520-2600 x2208 (Cell) 613-612-2708 [email protected] http://www.scs.carleton.ca/~arpwhite Expert Assessment of Stigmergy Abstract This report describes the current state of research in the area known as Swarm Intelligence. Swarm Intelligence relies upon stigmergic principles in order to solve complex problems using only simple agents. Swarm Intelligence has been receiving increasing attention over the last 10 years as a result of the acknowledgement of the success of social insect systems in solving complex problems without the need for central control or global information. In swarm- based problem solving, a solution emerges as a result of the collective action of the members of the swarm, often using principles of communication known as stigmergy. The individual behaviours of swarm members do not indicate the nature of the emergent collective behaviour and the solution process is generally very robust to the loss of individual swarm members. This report describes the general principles for swarm-based problem solving, the way in which stigmergy is employed, and presents a number of high level algorithms that have proven utility in solving hard optimization and control problems. Useful tools for the modelling and investigation of swarm-based systems are then briefly described.
  • Swarm Intelligence Among Humans the Case of Alcoholics

    Swarm Intelligence Among Humans the Case of Alcoholics

    Swarm Intelligence among Humans The Case of Alcoholics Andrew Schumann1 and Vadim Fris2 1Department of Cognitivistics, University of Information Technology and Management in Rzeszow, Sucharskiego 2, 35-225, Rzeszow, Poland 2Private Health Unitary Enterprise “Iscelenie”, Kazinca 120, 220000 Minsk, Belarus Keywords: Swarm Intelligence, Swarm Computing, Alcoholic, Illocutionary. Abstract: There are many forms of swarm behaviour, such as swarming of insects, flocking of birds, herding of quadrupeds, and schooling of fish. Sometimes people behave unconsciously and this behaviour of them has the same patterns as behaviour of swarms. For instance, pedestrians behave as herding or flocking, aircraft boarding passengers behave as ant colony, people in escape panic behave as flocking, etc. In this paper we propose a swarm model of people with an addictive behaviour. In particular, we consider small groups of alcohol-dependent people drinking together as swarms with a form of intelligence. In order to formalize this intelligence, we appeal to modal logics K and its modification K'. The logic K is used to formalize preference relation in the case of lateral inhibition in distributing people to drink jointly and the logic K' is used to formalize preference relation in the case of lateral activation in distributing people to drink jointly. 1 INTRODUCTION mammals, too, e.g. among naked mole-rats (Heterocephalus glaber sp.). In one colony they Usually, a social behaviour is understood as a have only one queen and one to three males to synonymous to a collective animal behaviour. It is reproduce, while other members of the colony are claimed that there are many forms of this behaviour just workers (Jarvis, 1981).
  • Science & Fiction-Febr, 2020 1 Off-Shore Fly Or Die

    Science & Fiction-Febr, 2020 1 Off-Shore Fly Or Die

    Science & Fiction-Febr, 2020 Off-Shore Fly or Die Holograms for Desert Locust Swarms Dr. Ugur Sevilmis Eastern Mediterranean Agricultural Research Institute, Adana, Turkey [email protected] Abstract A random walk search was conducted on academic papers to find a method for ecological combat of desert locust. Relatedness of papers and concepts were mainly analysed with the core idea of “Fighting Rooster” vs “Desert Locust “ interspecies competition. It was interesting to find a better preator than fighting roosters for locust swarms in the sea. Usage of a technology is required in this theoric approach to modify collective behavior of desert locusts which looks like the achilles heel of swarms. Keywords: Invasion, migration, predation, interspecies competition, süper-organism, swarm intelligence, collective intelligence, multi-species communities Conclusions A locust swarm is a highly adaptive süper-organism using collective behavior and collective intelligence. Complex interactions exists within its interiror and environmental relations coming from genetic evolution of this species. It is hard to totally change its capacity of foraging on herbage due to its high adaptation capacity. The best method to prevent it is to keep them away from terestial ecosystems and keep them off-shore to spend their fats. Just one day will be adequate to destroy a whole swarm; when the night comes. During night time when all individual visual sensors are shut down, the süper-organsim will be fragmented. Under dark and no landing condition, they will spend their energy. Dead bodies of locusts will be the landing zone for the rest of the huge swarm. But predator fishes will be vey happy to hunt 40 and 80 million locusts per square kilometer (a total of around 50 to 100 billion locusts per swarm, representing 100,000 to 200,000 tons of flesh).
  • An Equation of State for Insect Swarms Michael Sinhuber1,3,4, Kasper Van Der Vaart1,4, Yenchia Feng1, Andrew M

    An Equation of State for Insect Swarms Michael Sinhuber1,3,4, Kasper Van Der Vaart1,4, Yenchia Feng1, Andrew M

    www.nature.com/scientificreports OPEN An equation of state for insect swarms Michael Sinhuber1,3,4, Kasper van der Vaart1,4, Yenchia Feng1, Andrew M. Reynolds2 & Nicholas T. Ouellette1* Collective behaviour in focks, crowds, and swarms occurs throughout the biological world. Animal groups are generally assumed to be evolutionarily adapted to robustly achieve particular functions, so there is widespread interest in exploiting collective behaviour for bio-inspired engineering. However, this requires understanding the precise properties and function of groups, which remains a challenge. Here, we demonstrate that collective groups can be described in a thermodynamic framework. We defne an appropriate set of state variables and extract an equation of state for laboratory midge swarms. We then drive swarms through “thermodynamic” cycles via external stimuli, and show that our equation of state holds throughout. Our fndings demonstrate a new way of precisely quantifying the nature of collective groups and provide a cornerstone for potential future engineering design. Organisms on every size scale, from single-celled1 to highly complex2, regularly come together in groups. In many cases, such aggregations are collective, in that the group as a whole displays properties and functionality distinct from those of its individual members or simply their linear sum3,4. It is generally assumed that since evolution has led so many diferent kinds of animals to behave collectively, the performance of collective groups at whatever task they seek to achieve ought to be well beyond the capabilities of a single individual5, while also being robust to uncertain natural environments 6,7 and operating without the need for top-down control 8.
  • A Catalogue of Bacterial Swarm Behaviour

    A Catalogue of Bacterial Swarm Behaviour

    RESEARCH HIGHLIGHTS Nature Reviews Physics | https://doi.org/10.1038/s42254-020-0172- x | Published online 7 April 2020 ACTIVE MATTER A catalogue of bacterial swarm behaviour Even under adverse conditions length of its cells can be controlled such as starvation, some bacteria by known genetic manipulations. can efficiently expand and move This controllability enabled their colonies by rapidly migrating Be’er et al. to select between four en masse, a process known as cell aspect ratios for their colonies. Credit: Adapted from Be’er, A. et al. Commun. Phys. swarming. Although bacterial They recorded the motion of the 3, 66 (2020), CC BY 4.0 colonies have been studied by colonies on a surface for a range of physicists for years, there is not colony densities. Very sparse colonies of similar self-propelled rods with yet a complete picture of how the of B. subtilis do not move, whereas known interactions indicates that physical properties of the cells, very dense colonies are jammed. the bacterial behaviour is domina- such as their shape, govern the However, for intermediate densities, ted by short-range interactions. behaviour of the swarm. Now, cell aspect ratio plays a role in In contrast to long cells, shorter writing in Communications Physics, the colony’s behaviour. Longer cells swarm in clusters that have Avraham Be’er and colleagues fill cells form high-density clusters uniform density across the surface in another piece of the puzzle, of moving cells that are separated (lower panels of figure). Be’er et al. cataloguing the swarming behaviour by low- density regions containing posit that this uniform density of the rod-shaped Bacillus subtilis, only immobile cells (upper left arises because long-range hydro- as a function of the density of panel of figure).
  • Immune-Inspired Self-Healing Swarm Robotic Systems Amelia

    Immune-Inspired Self-Healing Swarm Robotic Systems Amelia

    Immune-Inspired Self-Healing Swarm Robotic Systems Amelia Ritahani Ismail Submitted for the degree of Doctor of Philosophy University of York Department of Computer Science September 2011 Abstract The field of artificial immune system (AIS) is an example of biologically inspired comput- ing that takes its inspiration from various aspects of immunology. Techniques from AIS have been applied in solving many different problems such as classification, optimisation and anomaly detection. However, despite the apparent success of the AIS approach, the unique advantages of AIS over and above other computational intelligence approaches are not clear. In order to address this, AIS practitioners need to carefully consider the application area and design methodologies that they adopt. It has been argued that of increasing importance is the development of a greater understanding of the underlying immunological system that acts as inspiration, as well as the understanding of the prob- lem that need to be solved before proposing the immune-inspired solution to solve the desired problem. This thesis therefore aims to pursue a more principled approach for the development of an AIS, considering the application areas that are suitable based on the underlying biological system under study, as well as the engineering problems that needs to be solved. This directs us to recognise a methodology for developing AIS that integrates several explicit modelling phases to extract the key features of the biological system. An analysis of the immunological literature acknowledges our immune inspira- tion: granuloma formation, which represents a chronic inflammatory reaction initiated by various infectious and non-infectious agents. Our first step in developing an AIS sup- ported by these properties is to construct an Unified Modelling Language (UML) model agent-based simulation to understand the underlying properties of granuloma formation.
  • Information Propagation in a Social Network: the Case of a Fish Schooling Algorithm

    Information Propagation in a Social Network: the Case of a Fish Schooling Algorithm

    Chapter 1 Information Propagation in a Social Network: The Case of a Fish Schooling Algorithm A. Brabazon, W. Cui and M. O’Neill Abstract The propagation of information about the environment amongst animals via social communication has attracted increasing research interest in recent decades with the realisation that many animal species engage in subtle forms of information transfer which had previously escaped notice. From an evolutionary perspective, the widespread existence of social communication mechanisms is not surprising given the significant benefits which can accrue to behaviours such as sharing of informa- tion on resources and on environmental threats. More generally, we can consider this process as information flowing between a network of nodes or agents, wherein each agent receives inputs from their senses and conspecifics, processes this information, and in turn through their resulting actions, can influence subsequent actions of other agents. Social communication mechanisms of organisms have inspired the develop- ment of several powerful families of optimisation algorithms including ant colony optimsation and honey bee optimisation algorithms. One interesting example of so- cial information propagation is provided by the shoaling and schooling behaviours of fish. In this chapter we develop an optimisation algorithm (the Fish Algorithm) which is inspired by the schooling behaviour of ‘golden shiner’ fish (Notemigonus Anthony Brabazon Complex Adaptive Systems Laboratory and School of Business University College Dublin Dublin, Ireland, e-mail: [email protected] Wei Cui Complex Adaptive Systems Laboratory and School of Business University College Dublin Dublin, Ireland, e-mail: [email protected] Michael O’Neill Complex Adaptive Systems Laboratory and School of Business University College Dublin Dublin, Ireland, e-mail: [email protected] 1 2 A.
  • Analysing Robot Swarm Behaviour Via Probabilistic Model Checking

    Analysing Robot Swarm Behaviour Via Probabilistic Model Checking

    Analysing Robot Swarm Behaviour via Probabilistic Model Checking Savas Konur, Clare Dixon, and Michael Fisher Department of Computer Science, University of Liverpool, UK {Konur, CLDixon, MFisher}@liverpool.ac.uk Published as Robotics and Autonomous Systems 60(2):199-213, Feb. 2012 Abstract An alternative to deploying a single robot of high complexity can be to utilize robot swarms comprising large numbers of identical, and much sim- pler, robots. Such swarms have been shown to be adaptable, fault-tolerant and widely applicable. However, designing individual robot algorithms to ensure effective and correct overall swarm behaviour is actually very diffi- cult. While mechanisms for assessing the effectiveness of any swarm algo- rithm before deployment are essential, such mechanisms have traditionally involved either computational simulations of swarm behaviour, or experi- ments with robot swarms themselves. However, such simulations or exper- iments cannot, by their nature, analyse all possible swarm behaviours. In this paper, we will develop and apply the use of automated probabilistic for- mal verification techniques to robot swarms, involving an exhaustive math- ematical analysis, in order to assess whether swarms will indeed behave as required. In particular we consider a foraging robot scenario to which we apply probabilistic model checking. 1 Introduction Robot swarms, comprising a number of simple and identical robots, have been de- veloped by many researchers and deployed in significant application areas [4, 5, 1]. Although the behaviour of each individual robot is simple, these robots work to- gether to achieve potentially quite complex swarm behaviour. Consequently, un- derstanding individual robot behaviour is easy, but predicting the overall swarm behaviour is much more difficult.
  • Classifying the Behaviors of Boid Swarms

    Classifying the Behaviors of Boid Swarms

    Classifying the Behaviors of Boid Swarms Zihao Liu, Ryan McArdle, Gianni Orlando 9 December 2020 Institute for Artificial Intelligence University of Georgia Athens, Georgia, USA [email protected], [email protected], [email protected] I.A BSTRACT force that is calculated against each of its neighbors. To achieve alignment, the boid calculates a force which will adjust First presented by Craig Reynolds in 1987, a boid is a class its velocity towards the average velocity of its neighbors. To of bird-like computational objects intended to simulate the achieve cohesion, the boid calculates a force that moves it flocking behavior that can be found in a number of animals. towards the centroid or average position of its neighbors. The While the concept of boids has been well established and behavior which emerges from an appropriate balance of these implemented for some time now in order to generate simu- forces allows a swarm of boids to simulate natural flocking lations of flocking swarms, the project of classifying swarms behavior in a tractable way, allowing one to define behavior of animals and classifying when they are exhibiting certain over the entire class rather than for each instance. The boid behaviors is a newer one, without as much success as of definition is also reliable in that it generates behaviors that yet. We utilize the Swarm Behavior Data Set uploaded by other distributive methods fail to achieve, such as allowing researchers at the University of New South Wales to the UCI a swarm to split to avoid colliding with an obstacle and Machine Learning Repository in order to train a classifier to recombine once they have passed the obstacle.
  • On the Role of Stigmergy in Cognition

    On the Role of Stigmergy in Cognition

    Noname manuscript No. (will be inserted by the editor) On the Role of Stigmergy in Cognition Lu´ısCorreia · Ana M. Sebasti~ao · Pedro Santana Received: date / Accepted: date Abstract Cognition in animals is produced by the self- 1 Introduction organized activity of mutually entrained body and brain. Given that stigmergy plays a major role in self-org- Swarm models are one of the most recent approaches anization of societies, we identify stigmergic behavior to cognition. They naturally map the cognitive capabil- in cognitive systems, as a common mechanism ranging ities of animal collectives such as termites as studied by from brain activity to social systems. We analyse natu- Turner (2011b), and they can also support individual ral societies and artificial systems exploiting stigmergy cognition, as shown in a realisation on mobile robots by to produce cognition. Several authors have identified Santana and Correia (2010). In the former case we are the importance of stigmergy in the behavior and cogni- in presence of cognition of the colony, which is situated tion of social systems. However, the perspective of stig- at a larger scale, when compared to that of each indi- mergy playing a central role in brain activity is novel, to vidual. In the latter case the swarm model forms the the best of our knowledge. We present several evidences perception of cognitive concepts by a single agent, the of such processes in the brain and discuss their impor- robot controller, which is external to the swarm. tance in the formation of cognition. With this we try Common to both cases above is the collective im- to motivate further research on stigmergy as a relevant pinging on the individuals' behavior and a strong em- component for intelligent systems.