Dissertation - HQ
Dissertation - HQ
Dissertation - HQ
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
122 Oceanography vs. behaviour<br />
Figure 6.2 Transition matrices for the simplest parameterisation of the onedimensional<br />
model. M 0 is the matrix associated with foraging (top), and M 1 is<br />
the matrix associated with swimming (bottom). In each case, lines are initial states,<br />
columns are final states, and elements of the matrix are transition probabilities<br />
(zeros are displayed as dots for clarity). States are indexed by energy reserves<br />
(E) and position (P)<br />
. . . and energy resources<br />
at recruitment<br />
scenario, interest is in strategies which maximise the probability of<br />
recruitment. The number of self-recruiting trajectories is potentially<br />
infinite, so maximising self-recruitment probability means selecting the<br />
strategues, hence trajectories along which survival is maximal. In other<br />
words, as self-recruitment is a prerequisite, the quantity optimised along<br />
recruiting trajectories (the gain) is, in fact, survival. As the introduction<br />
highlighted, this criterion is meaningful in terms of natural selection.<br />
However, more complex criteria can be specified, such as probability of<br />
return with maximum energy, with a given energy level, etc. The gain<br />
is then defined in terms of instantaneous gain (gain at each time step)<br />
and final gain (gain at the last time step).<br />
In this simple model, we focus on trajectories which optimise the<br />
probability of recruiting with maximum energy. This translates into zero