« Mathematics could stabilize peace treaties | Main | Mer forskning om bloggning »
december 17, 2003
Six Degrees of Immunization
Scientific Americal-artikeln 'Six Degrees of Immunization' Strategy Proposed berättar om en strategi för att vaccinera "superspridare" (super-spreaders) av virus.
Reuven Cohen of Bar-Ilan University in Israel and his colleagues note that random immunization programs require a large fraction of the population, typically 80 to 90 percent, be protected in order to stop the spread of disease. Alternatively, if enough information about the network and its connections is known, targeted immunization of the most highly connected individuals--so-called super-spreaders who have the potential to infect a high number of people--can be effective. Unfortunately, such information is difficult to acquire. The researchers instead propose a tactic known as acquaintance immunization. In it, a percentage of the population is selected at random and asked to identify a friend. Those friends, in turn, are vaccinated. According to the team's calculations, because super-spreaders know so many people, there is a high probability that they will be named at least once. As a result, immunization of a much smaller fraction of the population can successfully halt disease transmission. In addition, the authors note that their approach "can be used even before the epidemic starts spreading, since it does not require any knowledge of the chain of infection."
...
Cohen and his colleagues note that the technique is relevant to other types of networks, including terrorist ones.
Andra artiklar om detta:
Vaccinate Thy Neighbor
Uppdatering: Nature: Hub caps could cut vaccine costs
Papret som refereras är:
R. Cohen, D. ben-Avraham and S. Havlin
Efficient immunization strategies for computer networks and populations (PDF), Phys. Rev. Lett. 91, 247901 (2003)
Fler intressanta papers finns på Reuven Cohen publications-sida. Se även Media coverage.
Posted by hakank at december 17, 2003 09:02 FM Posted to Social Network Analysis/Complex Networks