WoTUG - The place for concurrent processes

BibTeX Proceedings details

BibTex html text dump of PaperDB database.

PaperDB is GPL Software, see: http://paperdb.sourceforge.net.

@InProceedings{SchallerMarshall03,
  title = "{A} {C}omparison of {H}igh {P}erformance, {P}arallel {C}omputing {J}ava {P}ackages",
  author= "Schaller, Nan C. and Marshall, Sidney W. and Cho, Yu-Fong",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "1--16",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "The high-performance computing community has developed
     numerous Java packages that support parallel and distributed
     computing. Most of these packages are designed for the
     typical parallel message passing and shared memory
     architectural paradigms. This paper presents the results of
     a recent study that included a web search for such packages,
     describes the paradigms implemented in them, and evaluates
     their performance on a parallel, 4-processor SMP machine
     using three benchmark programs that represent a mix of
     typical parallel applications, chosen from The Java Grande
     Benchmark Suite. A brief description of each package and a
     discussion its ease of installation and use are also
     provided."
}
@InProceedings{Petitpier03,
  title = "{A} {D}evelopment {M}ethod {B}oosted by {S}ynchronous {A}ctive {O}bjects",
  author= "Petitpierre, Claude",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "17--32",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "This paper presents a novel development method for
     interactive and distributedapplications. The benefit that
     this method provides is that the design gives clearguidance
     towards the implementation. The method is based on three
     mainelements: a concept of synchronous active objects that
     is closely related to thatfound in CSP; the Java
     environment; and a selection of the diagrams defined byUML.
     This approach alleviates many of the most serious problems
     that areencountered when using GUI builders, which hide the
     application structures andso make it difficult to devise
     sound architectures. The final part of the paper
     brieflydescribes an application generator that will help a
     developer to implement a designwhich is developed according
     to the approach advocated here."
}
@InProceedings{East03,
  title = "{T}utorial: {P}rioritised {S}ervice {A}rchitecture using {H}oneysuckle",
  author= "East, Ian R.",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "33--36",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "An update will be presented on the progress in establishing
     the Honeysuckleprogramming language [1] and its formal
     foundations. The latter are formallyaddressed in a paper
     currently under review for journal publication [2], but
     willbe summarised. They include formal definitions of
     service protocol plus servicenetwork/component (SNC) and the
     PSA design rule (PSADR), from which aproof of a priori
     deadlock-freedom emerges directly. Freedom from priority
     conflict(and thus inversion) is also easily guaranteed.
     Closure in the definition of systemand component guarantees
     true compositionality under both concurrency andprioritised
     alternation."
}
@InProceedings{ArrantssoVinter03,
  title = "{T}he {G}rid {B}lock {D}evice",
  author= "Arrantsson, Bardur and Vinter, Brian",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "37--48",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "In this paper we propose a distributed, replicated block
     device for The Gridbased on the the replication algorithm of
     Y. Amir[1]. The end goal is to supportapplication-specified
     semantics, both wrt. replication strategy and the
     consistencymodel used for reads and writes.We list the
     proposed features of the Grid Block Device and discuss
     various methodsof implementation."
}
@InProceedings{HappeVinter03,
  title = "{D}istributed {S}hared {M}emory in {G}lobal {A}rea {N}etworks",
  author= "Happe, Hans Henrik and Vinter, Brian",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "49--62",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "Distributed Shared Memory (DSM) has many advantages in
     heterogeneousenvironments, such as geographically distant
     clusters or The Grid. These includes:locality utilization
     and replication transparency. The fact that processes
     communicateindirectly through memory rather than directly,
     is giving DSM these advantages.This paper presents the
     design of Global PastSet (GPS) which is a DSM systemtargeted
     at global area networks. GPS is based on the DSM system
     PastSet [1] that hasbeen very effective in homogeneous
     cluster environments. GPS utilizes consistencycontrol
     migration and replication to scale in heterogeneous
     environments. This hasresulted in a token-based mutual
     exclusion algorithm that considers locality and analgorithm
     for locating replicas. GPS has been simulated in
     multi-cluster environmentswith up to 2048 nodes with very
     promising results."
}
@InProceedings{RajuRong03,
  title = "{A}utomatic {C}onversion of {CSP} to {CTJ}, {JCSP}, and {CCSP}",
  author= "Raju, V. and Rong, L. and Stiles, G. S.",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "63--81",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "We present tools that automatically convert a subset of
     machine-readable CSPscript to executable Java or C code. CSP
     is used to design and verify thecorrectness of large and
     complex systems of processes that interact only viaexplicit
     synchronous messages. These systems can be implemented in
     Javausing CTJ or JCSP, packages that add CSP-like features
     to Java, or in CCSP, apackage that adds similar features to
     standard C. Implementation of CSP systemscan be tedious and
     error-prone when large numbers of processes
     andcommunications are involved, and sorting out errors in
     channel naming or theordering of messages can be very
     time-consuming. The tools we have developedminimize such
     problems by converting the verified CSP descriptions
     ofcommunicating processes directly into Java or C code, thus
     guaranteeing thatchannels are correctly named and the
     communications occur in the proper order.This process can
     significantly cut development time."
}
@InProceedings{ONeillMoore03,
  title = "{A} {S}ingle {C}hip {S}olution for {D}istributed {P}rocessing {S}ystems",
  author= "O'Neill, Brian C. and Moore, P.W. and Clark, S.",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "83--90",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "This paper describes a processor and an inter-processor
     communications interfaceintegrated on a single chip for use
     in a distributed processing system. The system isbased on
     work of the electronic systems design and parallel
     processing group at theNottingham Trent University. The four
     main elements of the chip design areprocessor, memory,
     communication interface and packet routing switch all
     integratedunto one chip. This design is achieved by the use
     of the ALTERA ARM basedExcalibur system on a programmable
     chip (SOPC) containing an embedded processorand programmable
     logic. The paper describes the communication features
     andimplementation carried out by the research group to
     achieve this single chip processor."
}
@InProceedings{SmithHughes03,
  title = "{T}he denotational {S}emantics of {V}iew-{C}entric {R}easoning",
  author= "Smith, Marc L. and Hughes, Charles E. and Burke, Kyle W.",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "91--98",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "Both Lawrence’s HCSP [1] and Smith, et al’s VCR [2] (an
     earlier versionappears in [3]) extend CSP [4] with
     representations of truly concurrent events. Previously,VCR
     was described using an operational semantics, while the
     semantics ofHCSP’s Acceptances model, like those of the
     predominant CSP models described byRoscoe [5] (e.g., Traces,
     Failures / Divergences), are denotational. We now present
     adenotational semantics for VCRand, in so doing, propose an
     extension toHCSP (andpossibly other existing CSP models) to
     support View-Centric Reasoning. This workbrings VCR a step
     closer to being drawn within Hoare and He’s Unifying
     Theories ofProgramming [6] for further comparisons."
}
@InProceedings{Lawrence03,
  title = "{O}vertures and hesitant offers: hiding in {CSPP}",
  author= "Lawrence, Adrian E.",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "97--109",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "Hiding is an important and characteristic part of CSP.
     Defining it in thepresence of priority for CSPP needs care.
     The ideas of overtures and hesitant offersintroduced here
     arise naturally in the context of Acceptances. They provide
     clearinsight into the behaviour of hidden processes. And
     particularly illuminate the originof the nondeterminism
     which frequently arises from hiding.Keywords: CSP; CSPP;
     Denotational semantics; formal methods; concurrency;
     parallelsystems; occam; hardware compilation; priority;
     hiding."
}
@InProceedings{HilderinkBroenink03,
  title = "{S}ampling and timing a task for the environmental process",
  author= "Hilderink, Gerald H. and Broenink, Jan F.",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "111--124",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "Sampling and timing is considered a responsibility of the
     environment ofcontroller software. In this paper we will
     illustrate a concept whereby anenvironmental process and
     multi-way events play an important role in applyingtiming
     for untimed CSP software architectures. We use this timing
     concept forbuilding our control applications based on CSP
     concepts and with our CSP for C++(CTC++) library. We present
     a concept of sampling of control applications that
     isorthogonal to the application. This implies global timing
     on the basis of timedevents. We also support traditional
     local timing on the based of timed processes,"
}
@InProceedings{DullerPanesar03,
  title = "{P}arallel {P}rocessing - the pico{C}hip way!",
  author= "Duller, Andrew and Panesar, Gajinder and Towner, Daniel",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "125--138",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "This paper describes a new approach to parallel processing
     within thewell targetted application domain of wireless
     communications systems, using thepicoArrayTM. The
     picoArrayTM is a tiled-processor architecture, containing
     430 heterogeneousprocessors, connected through a novel,
     compile-time scheduled interconnect.We show how the features
     of the picoArrayTM allow deterministic processing tobe
     achieved, and how the tool chain allows programming to be
     performed effectivelyin a combination of high level assembly
     language and C. By handling a wide varietyof types of
     processing within the picoArrayTM a single design flow can
     be usedto produce complex communications systems. The
     effectiveness of this approach isdemonstrated through the
     use of the picoArrayTM to build a working 3G base-station."
}
@InProceedings{BrownWelch03,
  title = "{A}n {I}ntroduction to the {K}ent {C}++{CSP} {L}ibrary",
  author= "Brown, Neil C.C. and Welch, Peter H.",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "139--156",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "This paper reports on a CSP library for C++, developed over
     the past yearat the University of Kent. It is based on the
     OO-design and API of JCSP and thelightweight algorithms of
     KRoC occam, with extensions to exploit specific C++
     capabilities(such as templates). Both user-level and
     operating system threads are usedto provide a range of
     implementation options and semantics (e.g. for managing
     blockingsystem calls intelligently) that run efficiently
     under eitherWindows or Linux. Thelibrary is presented from
     the user’s point of view, mainly by way of a tutorial.
     Implementationdetails are also outlined and some benchmark
     results given. The performanceof C++CSP is between that of
     KRoC occam and JCSP — fairly close toKRoC."
}
@InProceedings{GonzalezBustacara03,
  title = "{A}gents for {C}oncurrent {P}rogramming",
  author= "Gonzalez, Enrique and Bustacara, Cesar and Avila, Jamir",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "157--166",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "This paper aims to demonstrate that concepts from
     Distributed Artificial Intelligence are very useful to
     design concurrent systems. The BESA framework and the AOPOA
     methodology are introduced as tools to achieve this goal.
     The Behavior-oriented, Event-driven and Social-based Agent
     (BESA) framework combines the concepts of MultiAgent Systems
     with the design of concurrent systems: an agent can be
     constructed as a set of behaviors; the notion of behaviors
     can be directly applied to concurrent systems design using
     the Agent Oriented Programming paradigm. The internal
     architecture of a BESA agent integrates two important
     features: a modular composition of behaviors and an event
     dispatcher based in a select like mechanism. The Agent
     Oriented Programming based in an Organizational Approach
     (AOPOA) methodology provides a systematic procedure to build
     complex system based in three concepts: a hierarchical
     recursive decomposition of the system, a goal-oriented role
     identification, and an evolution of the cooperation
     relationships linking the system components."
}
@InProceedings{NicoleEllis03,
  title = "occam for reliable embedded systems: lightweight runtime and model checking",
  author= "Nicole, Denis A. and Ellis, Sam and Hancock, Simon",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "167--172",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "We describe some more recent developments of the SPoC
     system. We describe a new module in the occam compiler which
     performs substantial simplifications of the run-time demands
     made by the compiled code. This has been used successfully
     both to target a simple PIC microcontroller and to generate
     input for the SMV model checker."
}
@InProceedings{CampbellStiles03,
  title = "{T}he {T}rebuchet",
  author= "Campbell, John and Stiles, G. S.",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "173--183",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "The Trebuchet is a hardware architecture for machines
     implementedfrom conventional software source programs. It is
     pseudo-asynchronous in that itdecouples the system clock
     from the computational logic, reducingelectromagnetic
     interference, smoothing current draw, and reducing
     pipelinelatency, key benefits of asynchronous designs.
     Performance of example programsis given."
}
@InProceedings{Boosten03,
  title = "{F}ormal {C}ontracts: {E}nabling {C}omponent {C}omposition",
  author= "Boosten, Marcel",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "185--197",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "Traditional component interaction is based on interface
     calls and callbacks. Such interaction can introduce
     integration faults, i.e., side effects at the moment of
     component integration. Solutions to such problems can be
     hard to apply, and may require drastic changes in the design
     of the involved components. This paper introduces Formal
     Contracts, a software construct that allows side-effect free
     component interaction, and thereby avoids the introduction
     of integration faults. Furthermore, via a state machine
     representing the inter-component contract, Formal Contracts,
     in addition to the static aspects, formally specify the
     dynamic aspects of component interaction. Formal Contracts
     are a pragmatic software mechanism that supports the full
     development cycle: from the specification and decomposition
     until the debugging, composition, and test of a system."
}
@InProceedings{SchweigleBarnes03,
  title = "{F}lexible, {T}ransparent and {D}ynamic occam {N}etworking {W}ith {KR}o{C}.net",
  author= "Schweigler, Mario and Barnes, Frederick R. M. and Welch, Peter H.",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "199--224",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "KRoC.net is an extension to KRoC supporting the distribution
     of occamchannels over networks, including the internet.
     Starting in 2001, the development ofKRoC.net has gone
     through a number of stages, each one making the system
     moreflexible, transparent and dynamic. It now enables the
     occam programmer to set upand close network channels
     dynamically. Configuration has been simplified. All
     occamPROTOCOLs can now be sent over network channels,
     without need for conversion.Many of the new dynamic features
     in occam have been used to improve KRoC.net.Many of the
     concepts in KRoC.net are similar to those in the JCSP
     Network Edition(JCSP.net), KRoC.net’s counterpart in the
     JCSP world. This paper will give an overviewover KRoC.net,
     its usage, its design and implementation, and its future. It
     willalso provide some benchmarks and discuss how the new
     occam features are beingused in the latest KRoC.net version."
}
@InProceedings{CronieHoeksema03,
  title = "{A} {CSP}-based {P}rocessing {A}rchitecture for a {F}lexible {MIMO}-{OFDM} {T}estbed",
  author= "Cronie, H. S. and Hoeksema, F. W. and Slump, C. H.",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "225--233",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "Future wireless communication systems require novel
     techniques to increasethe bitrate, coverage and mobility.
     One of these techniques is spatial multiplexingand we have
     investigated the use of a CSP-based kernel in the
     implementation ofa spatial multiplexing testbed. It turns
     out that the use of the CSP-based kernel notonly provides a
     good way of system modeling, but also provides a very
     scalable softwarearchitecture for the testbed. In future, we
     can change several system parameterswithout changing the
     software architecture. With the testbed we were able to
     verifythe concept of spatial multiplexing in an office
     environment."
}
@InProceedings{OrlicBroenink03,
  title = "{R}eal-time and fault tolerance in distributed control software",
  author= "Orlic, Bojan and Broenink, Jan F.",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "235--250",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "Closed loop control systems typically contain multitude of
     spatially distributed sensors and actuators operated
     simultaneously. So those systems are parallel and
     distributed in their essence. But mapping this parallelism
     onto the given distributed hardware architecture, brings in
     some additional requirements: safe multithreading, optimal
     process allocation, real-time scheduling of bus and network
     resources. Nowadays, fault tolerance methods and fast even
     online reconfiguration are becoming increasingly important.
     All those often conflicting requirements, make design and
     implementation of real-time distributed control systems an
     extremely difficult task, that requires substantial
     knowledge in several areas of control and computer science.
     Although many design methods have been proposed so far, none
     of them had succeeded to cover all important aspects of the
     problem at hand. [1] Continuous increase of production in
     embedded market, makes a simple and natural design
     methodology for real-time systems needed more then ever."
}
@InProceedings{Barnes03,
  title = "occwserv: {A}n occam {W}eb-{S}erver",
  author= "Barnes, Frederick R. M.",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "251--268",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "This paper presents ‘occwserv’, the occam web-server.
     This is a highlyconcurrent web-server, written in the occam
     multi-processing language, that supportsthe majority of the
     HTTP/1.1 protocol. Dynamic process and channel creation
     mechanismsare used to create scalable ‘server-farms’,
     each responsible for a particular webserverfunction - for
     example, reading client requests or running CGI processes.
     Thedesign of the web-server is presented, along with some
     early performance benchmarkresults. Although performance may
     appear a limiting factor (when compared to otherweb-servers
     such as Apache), much is gained from the simplicity and
     security of occam.Extending the web-server with new
     functionality, for example, is intuitive andlargely trivial
     - with the guarantees that code is free from race-hazard and
     aliasingerrors. An experimental non-standard addition, the
     OGI (occam Gateway Interface),is also presented. This
     provides a mechanism for dynamically loading and
     attachingpre-compiled occam processes to the running
     web-server, that can then handle oneor multiple client
     connections. A text-based style adventure game is examined
     briefly,that allows multiple clients to interact within a
     \"multi-user dungeon\" (MUD) styleenvironment."
}
@InProceedings{BarnesJacobson03,
  title = "{RM}o{X}: {A} raw-metal occam {E}xperiment",
  author= "Jacobsen, Christian L. and Barnes, Frederick R. M. and Vinter, Brian",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "269--288",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "Operating-systems are the core software component of many
     modern computersystems, ranging from small specialised
     embedded systems through to largedistributed
     operating-systems. This paper presents RMoX: a highly
     concurrent CSPbasedoperating-system written in occam. The
     motivation for this stems from theoverwhelming need for
     reliable, secure and scalable operating-systems. The
     majorityof operating-systems are written in C, a language
     that easily offers the level offlexibility required (for
     example, interfacing with assembly routines). C
     compilers,however, provide little or no mechanism to guard
     against race-hazard and aliasing errors,that can lead to
     catastrophic run-time failure (as well as to more subtle
     errors,such as security loop-holes). The RMoX
     operating-system presents a novel approachto
     operating-system design (although this is not the first
     CSP-based operating-system).Concurrency is utilised at all
     levels, resulting in a system design that is well
     defined,easily understood and scalable. The implementation,
     using the KRoC extended occam,provides guarantees of freedom
     from race-hazard and aliasing errors, and makesextensive use
     of the recently added support for dynamic process creation
     and channelmobility. Whilst targeted at mainstream
     computing, the ideas and methods presentedare equally
     applicable for small-scale embedded systems - where
     advantage can bemade of the lightweight nature of RMoX
     (providing fast interrupt responses, for example)."
}
@InProceedings{ArvindSoteloSal03,
  title = "{S}cheduling for {ILP} in the '{P}rocessor-as-a-{N}etwork'",
  author= "Arvind, D.K. and Sotelo-Salazar, S.",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "289--304",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "This paper explores the idea of the processor as an
     asynchronous network,called the micronet, of functional
     units which compute concurrently and
     communicateasynchronously. A micronet-based asynchronous
     processor exposes spatial as well astemporal concurrency. We
     analyse the performance of the
     ‘processor-as-a-network’by comparing three scheduling
     algorithms for exploiting Instruction Level
     Parallelism(ILP). Schedulers for synchronous architectures
     have relied on deterministic instructionexecution times. In
     contrast, ILP scheduling in micronet-based architectures is
     achallenge as it is less certain in advance when
     instructions start execution and whenresults become
     available. Performance results comparing the three
     schedulers are presentedfor SPEC95 benchmarks executing on a
     cycle-accurate model of the micronetarchitecture."
}
@InProceedings{Moore03,
  title = "{A}ccurate {C}alculation of {D}eme {S}izes for a {P}arallel {G}enetic {S}cheduling {A}lgorithm",
  author= "Moore, M.",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "305--313",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "The accuracies of three equations to determine the size of
     populations for serial and parallel genetic algorithms are
     evaluated when applied to a parallel genetic algorithm that
     schedules tasks on a cluster of computers connected via
     shared bus. This NP-complete problem is representative of a
     variety of optimisation problems for which genetic
     algorithms (GAs) have been shown to effectively approximate
     the optimal solution. However, empirical determination of
     parameters needed by both serial and parallel GAs is
     time-consuming, often impractically so in production
     environments. The ability to predetermine parameter values
     mathematically eliminates this difficulty. The parameter
     that exerts the most influence over the solution quality of
     a parallel genetic algorithm is the population size of the
     demes. Comparisons here show that the most accurate equation
     for the scheduling application is Cant\`{u}-Paz' serial
     population sizing calculation based on the gambler's ruin
     model [1]. The study presented below is part of an ongoing
     analysis of the effectiveness of parallel genetic algorithm
     parameter value computations based on schema theory. The
     study demonstrates that the correct deme size can be
     predetermined quantitatively for the scheduling problem
     presented here, and suggests that this may also be true for
     similar optimisation problems. This work is supported by
     NASA Grant NAG9-140."
}
@InProceedings{HilderinkJovanovic03,
  title = "{A} multimodal robotic control law modelled and implemented by the {CSP} - {GML}/{CT} framework",
  author= "Hilderink, Gerald H. and Jovanovic, Dusko S. and Broenink, Jan F.",
  editor= "Broenink, Jan F. and Hilderink, Gerald H.",
  pages = "315--334",
  booktitle= "{C}ommunicating {P}rocess {A}rchitectures 2003",
  isbn= "1 58603 381 6",
  year= "2003",
  month= "sep",
  abstract= "We use several formal methodologies for developing control
     applicationsat our Control Engineering research group. An
     important methodology we use fordesigning and implementing
     control software architecture is based on CSP concepts.These
     concepts allow us to glue multidisciplinary activities
     together and allow forformal stepwise refinement from design
     down to its implementation. This paperillustrates a
     trajectory and shows the usefulness of CSP diagrams for a
     simplemechatronic system. The simulation tool 20-SIM is used
     for creating the controllaws and our CTC++ package is used
     for coding in C++."
}

If you have any comments on this database, including inaccuracies, requests to remove or add information, or suggestions for improvement, the WoTUG web team are happy to hear of them. We will do our best to resolve problems to everyone's satisfaction.

Copyright for the papers presented in this database normally resides with the authors; please contact them directly for more information. Addresses are normally presented in the full paper.

Pages © WoTUG, or the indicated author. All Rights Reserved.
Comments on these web pages should be addressed to: www at wotug.org

Valid HTML 4.01!