Now showing 1 - 3 of 3
  • Publication
    Evolution with state-dependent mutations
    (University College Dublin. School of Economics, 1994-08) ;
    Recent evolutionary models have introduced "small mutation rates" as a way of refining predictions of long-run behavior. We show that if mutation rates are allowed to vary across states, then mutations no longer narrow the set of possible preditions. In particular, given any model of the effect of mutations, any invariant distribution of the "mutationless" process is close to an invariant distribution of the process with appropriately chosen small mutation rates.
      698
  • Publication
    Optimal patent length
    (University College Dublin. Geary Institute, 2008-03-18)
    The intent of the patent system is to encourage innovation by granting the innovator exclusive rights to a discovery for a limited period of time: with monopoly power, the innovator can recover the costs of creating the innovation which otherwise might not have existed. And, over time, the resulting innovation makes everyone better off. This presumption of improved social welfare is considered here. The paper examines the impact of patents on welfare in an environment where there are large numbers of (small) innovators — such as the software industry. With patents, because there is monopoly for a limited time the outcome is necessarily not socially optimal, although social welfare may be higher than in the no-patent state. Patent acquisition and ownership creates two opposing incentives at the same time: the incentive to acquire monopoly rights conferred by the patent spurs innovation, but subsequent ownership of those rights inhibits innovation (both own innovation and that of others). On balance, which effect will dominate? In the framework of this paper separate circumstances are identified under which patents are either beneficial or detrimental to innovation and welfare; and comparisons are drawn with the socially optimal level of investment in innovation.
      427
  • Publication
    Player type distributions as state variables and information revelation in zero sum repeated games with discounting
    (Institute for Operations Research and the Management Sciences (INFORMS), 1992-08)
    This paper examines the role of the player type distributions in repeated zero sum games of incomplete information with discounting of payoffs. In particular the strategic "sufficiency" of the posterior distributions for histories and the Limiting properties of the posterior sequence are discussed. It is shown that differentiability of the value function is sufficient to allow the posteriors to serve as "state" variables for histories. The limiting properties of the posterior distributions are considered and a characterization given of the set of possible limit points of the posterior distribution. This characterization is given in terms of the "value" of information in the one-stage game.
      449