“We con­sider the prob­lem of sched­ul­ing in mul­ti­hop wire­less net­works sub­ject to inter­fer­ence con­straints. We con­sider a graph based rep­re­sen­ta­tion of wire­less net­works, where sched­uled links adhere to the K-​​hop link inter­fer­ence model. We develop a dis­trib­uted greedy heuris­tic for this sched­ul­ing prob­lem. Fur­ther, we show that this dis­trib­uted greedy heuris­tic com­putes the exact same sched­ule as the cen­tral­ized greedy heuristic.”

“…Here we study the quan­ti­ta­tive rela­tion between adap­tive response and back­ground com­pen­sa­tion within a mod­el­ing frame­work. In con­trast to the com­monly held view, we show that any par­tic­u­lar type of adap­tive response is nei­ther suf­fi­cient nor nec­es­sary for adap­tive enlarge­ment of dynamic range. In par­tic­u­lar a pre­cise adap­tive response, where sys­tem activ­ity is main­tained at a con­stant level at steady state, does not ensure a large dynamic range nei­ther in input sig­nal nor in sys­tem out­put. A gen­eral mech­a­nism for input dynamic range enlarge­ment comes about from the activity-​​dependent mod­u­la­tion of pro­tein respon­sive­ness by mul­ti­ple bio­chem­i­cal mod­i­fi­ca­tion, regard­less of the type of adap­tive response it induces. There­fore hier­ar­chi­cal bio­chem­i­cal processes such as methy­la­tion and phos­pho­ry­la­tion are nat­ural can­di­dates to induce this prop­erty in sig­nalling systems.”

“Many mod­els of mar­ket dynam­ics make use of the idea of wealth exchanges among eco­nomic agents. A sim­ple anal­ogy com­pares the wealth in a soci­ety with the energy in a phys­i­cal sys­tem, and the trade between agents to the energy exchange between mol­e­cules dur­ing col­li­sions. How­ever, while in phys­i­cal sys­tems the equipar­ti­tion of energy is valid, in most exchange mod­els for eco­nomic mar­kets the sys­tem con­verges to a very unequal “con­densed” state, where one or a few agents con­cen­trate all the wealth of the soci­ety and the wide major­ity of agents shares zero or a very tiny frac­tion of the wealth. Here we present an exchange model where the goal is not only to avoid con­den­sa­tion but also to reduce the inequal­ity; to carry out this objec­tive the choice of inter­act­ing agents is not at ran­dom, but fol­lows an extremal dynam­ics reg­u­lated by the wealth of the agent.…”

“One direc­tion for future research would be to investi– gate to what extent these coun­terex­am­ples are spe­cial. For exam­ple, are all shapes which repel a point charge sim­i­lar to the hemi­sphere geom­e­try dis­cussed here, or are there com­pletely dif­fer­ent kinds of geome­tries with this prop­erty? More specif­i­cally, is it pos­si­ble to achieve re– pul­sion with a con­vex metal­lic object? One can ask sim– ilar ques­tions about Casimir repul­sion. There are many open ques­tions here—we have only just begun to under– stand these coun­ter­in­tu­itive geo­met­ric effects.”

“Rotor-​​router net­works are dis­crete ana­logues of con­tin­u­ous lin­ear sys­tems such as elec­tri­cal cir­cuits; they are also deter– min­is­tic ana­logues of sto­chas­tic sys­tems such as ran­dom walk processes. These analo­gies per­mit one to design rotor-​​router net­works to com­pute numer­i­cal quan­ti­ties asso­ci­ated with lin– ear and/​or sto­chas­tic sys­tems. These dis­trib­uted com­pu­ta­tions can behave sta­bly even in the pres­ence of sig­nif­i­cant disruption.”

“Behavior-​​Driven Devel­op­ment (BDD) is a spec­i­fi­ca­tion tech­nique that auto­mat­i­cally cer­ti­fies that all func­tional require­ments are treated prop­erly by source code, through the con­nec­tion of the tex­tual descrip­tion of these require­ments to auto­mated tests. Given that in some areas, in spe­cial Enter­prise Infor­ma­tion Sys­tems, require­ments are iden­ti­fied by Busi­ness Process Mod­el­ing — which uses graph­i­cal nota­tions of the under­ly­ing busi­ness processes, this paper aims to pro­vide a map­ping from the basic con­structs that form the most com­mon BPM lan­guages to Behav­ior Dri­ven Devel­op­ment constructs.”

“The study of net­works has grown into a sub­stan­tial inter­dis­ci­pli­nary endeavor across the nat­ural, social, and infor­ma­tion sci­ences. Yet there have been very few attempts to inves­ti­gate the inter­re­lat­ed­ness of the dif­fer­ent classes of net­works stud­ied by dif­fer­ent dis­ci­plines. Here, we intro­duced a frame­work to estab­lish a tax­on­omy of net­works from var­i­ous ori­gins. The pro­vi­sion of this fam­ily tree not only helps under­stand the kin­ship of net­works, but also facil­i­tates the trans­fer of empir­i­cal analy­sis, the­o­ret­i­cal mod­el­ing, and con­cep­tual devel­op­ments across dis­ci­pli­nary bound­aries. The frame­work is based on prob­ing the meso­scopic prop­er­ties of net­works, an impor­tant source of het­ero­gene­ity for their struc­ture and func­tion. Using our method, we com­puted a tax­on­omy for 752 indi­vid­ual net­works and a sep­a­rate tax­on­omy for 12 net­work classes. We also com­puted three within-​​class tax­onomies for polit­i­cal, fun­gal, and finan­cial net­works, and found them to be insight­ful in each case.”

“…The h index is com­pared with the Degree cen­tral­ity (a local mea­sure), the Between­ness and Eigen­vec­tor cen­tral­i­ties (two non-​​local mea­sures) in the case of a bio­log­i­cal net­work (Yeast inter­ac­tion protein-​​protein net­work) and a lin­guis­tic net­work (Moby The­saurus II). In both net­works, the Hirsch index has poor cor­re­la­tion with Between­ness cen­tral­ity but cor­re­lates well with Eigen­vec­tor cen­tral­ity, spe­cially for the more impor­tant nodes that are rel­e­vant for rank­ing pur­poses, say in Search Engine Opti­miza­tion. In the the­saurus net­work, the h index seems even to out­per­form the Eigen­vec­tor cen­tral­ity mea­sure as eval­u­ated by sim­ple lin­guis­tic criteria.”

“Despite the avail­abil­ity of very detailed data on finan­cial mar­ket, agent-​​based mod­el­ing is hin­dered by the lack of infor­ma­tion about real trader behav­ior. This makes it impos­si­ble to val­i­date agent-​​based mod­els, which are thus reverse-​​engineering attempts. This work is a con­tri­bu­tion to the build­ing of a set of styl­ized facts about the traders them­selves. Using the client data­base of Swis­squote Bank SA, the largest on-​​line Swiss bro­ker, we find empir­i­cal rela­tion­ships between turnover, account val­ues and the num­ber of assets in which a trader is invested. A the­ory based on sim­ple mean-​​variance port­fo­lio opti­miza­tion that cru­cially includes vari­able trans­ac­tion costs is able to repro­duce faith­fully the observed behav­iors. We finally argue that our results bring into light the col­lec­tive abil­ity of a pop­u­la­tion to con­struct a mean-​​variance port­fo­lio that takes into account the struc­ture of trans­ac­tion costs.”

“We con­sider the prob­lem of para­me­ter esti­ma­tion for a sys­tem of ordi­nary dif­fer­en­tial equa­tions from noisy obser­va­tions on a solu­tion of the sys­tem. In case the sys­tem is non­lin­ear, as it typ­i­cally is in prac­ti­cal appli­ca­tions, an ana­lytic solu­tion to it usu­ally does not exist. Con­se­quently, straight­for­ward esti­ma­tion meth­ods like the ordi­nary least squares method depend on repet­i­tive use of numer­i­cal inte­gra­tion in order to deter­mine the solu­tion of the sys­tem for each of the para­me­ter val­ues con­sid­ered, and to find sub­se­quently the para­me­ter esti­mate that min­imises the objec­tive func­tion. This induces a huge com­pu­ta­tional load to such esti­ma­tion meth­ods. We pro­pose an esti­ma­tor that is defined as a min­imiser of an appro­pri­ate dis­tance between a non­para­met­ri­cally esti­mated deriv­a­tive of the solu­tion and the right-​​hand side of the sys­tem applied to a non­para­met­ri­cally esti­mated solution.…”

“We present designs of 2D isotropic, dis­or­dered pho­tonic mate­ri­als of arbi­trary size with com­plete band gaps block­ing all direc­tions and polar­iza­tions. The designs with the largest gaps are obtained by a con­strained opti­miza­tion method that starts from a hype­r­uni­form dis­or­dered point pat­tern, an array of points whose num­ber vari­ance within a spher­i­cal sam­pling win­dow grows more slowly than the vol­ume. We argue that hype­r­uni­for­mity, com­bined with uni­form local topol­ogy and short-​​range geo­met­ric order, can explain how com­plete pho­tonic band gaps are pos­si­ble with­out long-​​range trans­la­tional order. We note the ram­i­fi­ca­tions for elec­tronic and phononic band gaps in dis­or­dered materials.”

“Sys­tems whose orga­ni­za­tion dis­plays causal asym­me­try con­straints, from evo­lu­tion­ary trees to river basins or trans­port net­works, can be often described in terms of directed paths (causal flows) on a dis­crete state space. Such a set of paths defines a feed-​​forward, acyclic net­work. A key prob­lem asso­ci­ated with these sys­tems involves char­ac­ter­iz­ing their intrin­sic degree of path reversibil­ity: given an end node in the graph, what is the uncer­tainty of recov­er­ing the process back­wards until the ori­gin? Here we pro­pose a novel con­cept, \textit{topological reversibil­ity}, which rig­or­ously weigths such uncer­tainty in path depen­dency quan­ti­fied as the min­i­mum amount of infor­ma­tion required to suc­cess­fully revert a causal path.…”

“In this paper, a para­met­ric level set method for recon­struc­tion of obsta­cles in gen­eral inverse prob­lems is con­sid­ered. Gen­eral evo­lu­tion equa­tions for the recon­struc­tion of unknown obsta­cles are derived in terms of the under­ly­ing level set para­me­ters. We show that using the appro­pri­ate form of para­me­ter­iz­ing the level set func­tion results a sig­nif­i­cantly lower dimen­sional prob­lem, which bypasses many dif­fi­cul­ties with tra­di­tional level set meth­ods, such as reg­u­lar­iza­tion, re-​​initialization and use of signed dis­tance function.…”

“… The aim is to find effi­cient decom­po­si­tions that simul­ta­ne­ously min­i­mize the irra­di­a­tion time, the car­di­nal­ity of the decom­po­si­tion and the setup-​​time to con­fig­ure the multi-​​leaf col­li­ma­tor at each step of the decom­po­si­tion. We pro­pose for this NP-​​hard mul­ti­ob­jec­tive com­bi­na­to­r­ial prob­lem a heuris­tic, based on the adap­ta­tion of the two-​​phase Pareto local search. Exper­i­ments are car­ried out on dif­fer­ent size instances and the results are reported.”

“The knap­sack prob­lem (KP) and its mul­ti­di­men­sional ver­sion (MKP) are basic prob­lems in com­bi­na­to­r­ial opti­miza­tion. In this paper we con­sider their mul­ti­ob­jec­tive exten­sion (MOKP and MOMKP), for which the aim is to obtain or to approx­i­mate the set of effi­cient solu­tions. In a first step, we clas­sify and describe briefly the exist­ing works, that are essen­tially based on the use of meta­heuris­tics. In a sec­ond step, we pro­pose the adap­ta­tion of the two-​​phase Pareto local search (2PPLS) to the res­o­lu­tion of the MOMKP. With this aim, we use a very-​​large scale neigh­bor­hood (VLSN) in the sec­ond phase of the method, that is the Pareto local search. We com­pare our results to state-​​of-​​the-​​art results and we show that we obtain results never reached before by heuris­tics, for the biob­jec­tive instances. Finally we con­sider the exten­sion to three-​​objective instances.”

“We con­sider the prob­lem of com­put­ing a response curve for binary cel­lu­lar automata — that is, the curve describ­ing the depen­dence of the den­sity of ones after many iter­a­tions of the rule on the ini­tial den­sity of ones. We demon­strate how this prob­lem could be approached using rule 130 as an exam­ple. For this rule, preim­age sets of finite strings exhibit rec­og­niz­able pat­terns, and it is there­fore pos­si­ble to com­pute both car­di­nal­i­ties of preim­ages of cer­tain finite strings and prob­a­bil­i­ties of occur­rence of these strings in a con­fig­u­ra­tion obtained by iter­at­ing a ran­dom ini­tial con­fig­u­ra­tion $n$ times. Response curves can be rig­or­ously cal­cu­lated in both one– and two-​​dimensional ver­sions of CA rule 130. We also dis­cuss a spe­cial case of totally dis­or­dered ini­tial con­fig­u­ra­tions, that is, ran­dom con­fig­u­ra­tions where the den­sity of ones and zeros are equal to ¹⁄₂.”

“This paper presents a new numer­i­cal approach to the study of non-​​periodicity in sig­nals, which can com­ple­ment the max­i­mal Lya­punov expo­nent method for deter­min­ing chaos tran­si­tions of a given dynam­i­cal sys­tem. The pro­posed tech­nique is based on the con­tin­u­ous wavelet trans­form and the wavelet mul­tires­o­lu­tion analy­sis. A new para­me­ter, the \textit{scale index}, is intro­duced and inter­preted as a mea­sure of the degree of the signal’s non-​​periodicity. This method­ol­ogy is suc­cess­fully applied to three clas­si­cal dynam­i­cal sys­tems: the Bonhoeffer-​​van der Pol oscil­la­tor, the logis­tic map, and the Henon map.”

“Fea­tur­ing auto­matic Walk-​​Forward Test­ing, Multi-​​monitor float­ing charts, sym­bol and inter­val link­ing, drag-​​and-​​drop indi­ca­tor cre­ation, Indus­try fastest, Unlimited-​​symbol True Portfolio-​​Level Back­test­ing and Opti­miza­tion, now with Smart Evo­lu­tion­ary algo­rithms, scal­ing, market-​​neutral sys­tem sup­port and mul­ti­ple cur­rency han­dling, free Fun­da­men­tal data, Mul­ti­ple Time-​​Frame sup­port, 3D opti­miza­tion charts, new Account man­ager, auto­mated trad­ing inter­face, vol­ume pro­file, object-​​oriented chart­ing, draw­ing lay­ers, multi-​​window lay­outs, formula-​​based alerts, easy-​​to-​​use for­mula edi­tor, equity func­tion, unique com­pos­ite indi­ca­tors, built-​​in web research browser, direct link to eSig­nal, Inter­ac­tive Bro­kers, IQFeed, myTrack, Fast­Track, QP2, TC2000, any DDE com­pli­ant feed, MS and more…”

“Hugin has now reached sta­ble state: the soft­ware is rec­om­mended for gen­eral use.”

“I’ve found there are two types of thoughts espe­cially worth avoiding—thoughts like the Nile Perch in the way they push out more inter­est­ing ideas. One I’ve already men­tioned: thoughts about money. Get­ting money is almost by def­i­n­i­tion an atten­tion sink. The other is dis­putes. These too are engag­ing in the wrong way: they have the same velcro-​​like shape as gen­uinely inter­est­ing ideas, but with­out the sub­stance. So avoid dis­putes if you want to get real work done.”

“Recov­er­ing the 3D struc­ture of the scene from images yields use­ful infor­ma­tion for tasks such as shape and scene recog­ni­tion, object detec­tion, or motion plan­ning and object grasp­ing in robot­ics. In this the­sis, we intro­duce a gen­eral machine learn­ing approach called unsu­per­vised CRF learn­ing based on max­i­miz­ing the con­di­tional like­li­hood. We apply our approach to com­puter vision sys­tems that recover the 3-​​D scene geom­e­try from images. We focus on recov­er­ing 3D geom­e­try from sin­gle images, stereo pairs and video sequences. Build­ing these sys­tems requires algo­rithms for doing infer­ence as well as learn­ing the para­me­ters of con­di­tional Markov ran­dom fields (MRF). Our sys­tem is trained unsu­per­vis­edly with­out using ground-​​truth labeled data.…”

“We study dis­trib­uted opti­miza­tion in net­worked sys­tems, where nodes coop­er­ate to find the opti­mal quan­tity of com­mon inter­est, x = x^\star. The objec­tive func­tion of the cor­re­spond­ing opti­miza­tion prob­lem is the sum of pri­vate (known only by a node,) con­vex, nodes’ objec­tives and each node imposes a pri­vate con­vex con­straint on the allowed val­ues of x. We solve this prob­lem for generic con­nected net­work topolo­gies with asym­met­ric ran­dom link fail­ures with a novel dis­trib­uted, decen­tral­ized algorithm.”

“The pres­ence of new kinds of investors (with moti­va­tions and objec­tives dif­fer­ent from the hedgers and spec­u­la­tors) has altered com­mod­ity mar­kets. The lat­ter are now sub­ject to forces dif­fer­ent from those in the peri­ods cov­ered by the research stud­ies. The actions of pas­sive index investors are new fac­tors impact­ing the prices of com­mod­ity futures. I wouldn’t nec­es­sar­ily expect com­modi­ties to con­tinue per­form­ing as they may have per­formed in the decades before the 2000s – i.e. reg­is­ter­ing aver­age returns sim­i­lar to stocks at lower lev­els of risk and with low cor­re­la­tions to stocks.”

“HTSQL (“Hyper Text Struc­tured Query Lan­guage”) is a schema-​​driven URI-​​to-​​SQL trans­la­tor that takes a request over HTTP, con­verts it to a SQL query, exe­cutes the query against a data­base, and returns the results in a for­mat best suited for the user agent (CSV, HTML, etc.).
HTSQL 2.0 is a work in progress and not yet ready for pro­duc­tion use. The ini­tial sup­ported release will be in Octo­ber of 2010.”

“Select the his­tor­i­cal mar­ket data prod­ucts below
Here you can select the prod­ucts you are inter­ested in. Click on the product’s name to find out more about it. Press the Con­tinue but­ton to place an order or to get a quote.”