These are my recent Pin​board​.in links:

• [1201.5440] Self-​​assembly of anisotropic soft par­ti­cles in two dimensions

  “The self assem­bly of core-​​corona discs inter­act­ing via anisotropic poten­tials is inves­ti­gated using Monte Carlo com­puter sim­u­la­tions. A min­i­mal inter­ac­tion poten­tial that incor­po­rates anisotropy in a sim­ple way is intro­duced. It con­sists in a core-​​corona archi­tec­ture in which the cen­ter of the core is shifted with respect to the cen­ter of the corona. Anisotropy can thus be tuned by pro­gres­sively shift­ing the posi­tion of the core. Despite its sim­plic­ity, the sys­tem self orga­nize in a rich vari­ety of struc­tures includ­ing stripes, tri­an­gu­lar and rec­tan­gu­lar lat­tices, and unusual plas­tic crys­tals. Our results indi­cate that the amount of anisotropy does not alter the lat­tice spac­ing and only influ­ences the type of clus­ter­ing (stripes, micells, etc.) of the indi­vid­ual particles.”
  
  self-​​assembly biologically-​​inspired sim­u­la­tion pattern-​​formation condensed-​​matter

• [1201.5477] Entropy-​​growth-​​based model of emo­tion­ally charged online dialogues

  “We ana­lyze emo­tion­ally anno­tated mas­sive data from IRC (Inter­net Relay Chat) and model the dia­logues between its par­tic­i­pants by assum­ing that the dri­ving force for the dis­cus­sion is the entropy growth of emo­tional prob­a­bil­ity dis­tri­b­u­tion. This process is claimed to be cor­re­lated to the emer­gence of the power-​​law dis­tri­b­u­tion of the dis­cus­sion lengths observed in the dia­logues. We per­form numer­i­cal sim­u­la­tions based on the noticed phe­nom­e­non obtain­ing a good agree­ment with the real data. Finally, we pro­pose a method to arti­fi­cially pro­long the dura­tion of the dis­cus­sion that relies on the entropy of emo­tional prob­a­bil­ity distribution.”
  
  oh-​​look-​​power-​​laws flame-​​wars social-​​dynamics com­plex­ol­ogy cultural-​​dynamics

• [1201.4955] Coor­di­na­tion, Dif­fer­en­ti­a­tion and Fair­ness in a pop­u­la­tion of coop­er­at­ing agents

  “In a recent paper, we ana­lyzed the self-​​assembly of a com­plex coop­er­a­tion net­work. The net­work was shown to approach a state, where every agent invests the same amount of resources. Nev­er­the­less, highly-​​connected agents arise that extract extra-​​ordinarily high pay­offs while con­tribut­ing com­pa­ra­bly lit­tle to any of their coop­er­a­tions. Here, we inves­ti­gate a vari­ant of the model, in which highly-​​connected agents have access to addi­tional resources. We study ana­lyt­i­cally and numer­i­cally whether these resources are invested in exist­ing col­lab­o­ra­tions, lead­ing to a fairer load dis­tri­b­u­tion, or in estab­lish­ing new col­lab­o­ra­tions, lead­ing to an even less fair dis­tri­b­u­tion of loads and payoffs.”
  
  col­lab­o­ra­tion social-​​capital agent-​​based network-​​theory com­plex­ol­ogy nudge-​​targets

• [1201.5426] Con­straint Prop­a­ga­tion as Infor­ma­tion Maximization

  “Dana Scott used the par­tial order among par­tial func­tions for his math­e­mat­i­cal model of recur­sively defined func­tions. He inter­preted the par­tial order as one of infor­ma­tion con­tent. In this paper we elab­o­rate on Scott’s sug­ges­tion of regard­ing com­pu­ta­tion as a process of infor­ma­tion max­i­miza­tion by apply­ing it to the solu­tion of con­straint sat­is­fac­tion prob­lems. Here the method of con­straint prop­a­ga­tion can be inter­preted as decreas­ing uncer­tainty about the solu­tion — that is, as gain in infor­ma­tion about the solu­tion. As illus­tra­tive exam­ple we choose numer­i­cal con­straint sat­is­fac­tion prob­lems to be solved by inter­val con­straints. To facil­i­tate this approach to con­straint solv­ing we for­mu­late con­straint sat­is­fac­tion prob­lems as for­mu­las in pred­i­cate logic. This neces­si­tates extend­ing the usual seman­tics for pred­i­cate logic so that mean­ing is assigned not only to sen­tences but also to for­mu­las with free variables.”
  
  computer-​​science quite-​​interesting constraint-​​processing computational-​​methods

• [1201.4459] An effi­cient par­al­lel algo­rithm for the longest path prob­lem in meshes

  “In this paper, first we give a sequen­tial linear-​​time algo­rithm for the longest path prob­lem in meshes. This algo­rithm can be con­sid­ered as an improve­ment of [13]. Then based on this sequen­tial algo­rithm, we present a constant-​​time par­al­lel algo­rithm for the prob­lem which can be run on every par­al­lel machine.”
  
  algo­rithms graph-​​theory computational-​​complexity nudge-​​targets

• [1201.4417] Insta­bil­i­ties and Pat­terns in Cou­pled Reaction-​​Diffusion Layers

  “We study insta­bil­i­ties and pat­tern for­ma­tion in reaction-​​diffusion lay­ers that are dif­fu­sively cou­pled. For two-​​layer sys­tems of iden­ti­cal two-​​component reac­tions, we ana­lyze the sta­bil­ity of homo­ge­neous steady states by exploit­ing the block sym­met­ric struc­ture of the lin­ear prob­lem. There are eight pos­si­ble pri­mary bifur­ca­tion sce­nar­ios, includ­ing a Turing-​​Turing bifur­ca­tion that involves two dis­parate length scales whose ratio may be tuned via the inter-​​layer cou­pling. For sys­tems of $n$-component lay­ers and non-​​identical lay­ers, the lin­ear problem’s block form allows approx­i­mate decom­po­si­tion into lower-​​dimensional lin­ear prob­lems if the cou­pling is suf­fi­ciently weak. As an exam­ple, we apply these results to a two-​​layer Brus­se­la­tor sys­tem. The com­pet­ing length scales engi­neered within the lin­ear prob­lem are read­ily appar­ent in numer­i­cal sim­u­la­tions of the full sys­tem. Select­ing a $sqrt{2}$:1 length scale ratio pro­duces an unusual steady square pattern.”
  
  cute emergent-​​design pattern-​​formation com­plex­ol­ogy nudge-​​targets nonlinear-​​dynamics

• [1201.4737] Pro­duc­tion Sys­tem Rules as Pro­tein Com­plexes from Genetic Reg­u­la­tory Networks

  “This short paper intro­duces a new way by which to design pro­duc­tion sys­tem rules. An indi­rect encod­ing scheme is pre­sented which views such rules as pro­tein com­plexes pro­duced by the tem­po­ral behav­iour of an arti­fi­cial genetic reg­u­la­tory net­work. This ini­tial study begins by using a sim­ple Boolean reg­u­la­tory net­work to pro­duce tra­di­tional ternary-​​encoded rules before mov­ing to a fuzzy vari­ant to pro­duce real-​​valued rules. Com­pet­i­tive per­for­mance is shown with related genetic reg­u­la­tory net­works and rule-​​based sys­tems on bench­mark problems.”
  
  evolutionary-​​algorithms production-​​systems computer-​​science emergent-​​design