362
M. Antoniou et al.
– In general, larger offspring size speeds up the search process and increases the
diversity of solutions (in earlier stages of the optimisation process).
• Mutation rate or mutation probability
– Also referred as p m .
– The frequency of new mutations in a single gene (variable representation) or
individual (solution representation) over time.
– In general, higher mutation rate increases the versatility of the solutions within
the population, but it also increases the amount of randomness in the search.
• Mutation step size
– Also referred as σ .
– The size of the change of mutation rate compared to its previous value. In the
case of real-valued search spaces, mutation is usually performed by adding a
normally distributed random value to each optimised variable.
– In general, higher mutation step size has a similar influence as the higher
mutation rate.
• Crossover rate or crossover probability
– Also referred as p c .
– The probability that crossover is performed between two chosen solutions.
– In general, lower crossover probability prevents exchange of genetic material
and causes slower convergence.
• Selection pressure
– It is the degree to which the better individuals are favoured. Which solutions
will provide the individual with an increased chance of surviving over others,
i.e. to be forced to contribute to the next generation. Therefore, individuals
with certain phenotypes have an advantage when it comes to survival and
reproduction.
– The higher the selection pressure, the more good solutions are kept from
generation to generation. It is related to elitism strategy, which ensures good
solutions to proceed to the next generation.
• Tournament size
– Tournament selection involves running tournaments amongst pairs (or groups)
of individuals chosen at random from the population. The winner of each
tournament (the one with the best fitness) is selected for the next phase (i.e.
crossover).
– When tournament size is higher, the weaker solutions have a smaller chance
to be selected.
• Number of generations
– Also referred as n g .
M. Antoniou et al.
– In general, larger offspring size speeds up the search process and increases the
diversity of solutions (in earlier stages of the optimisation process).
• Mutation rate or mutation probability
– Also referred as p m .
– The frequency of new mutations in a single gene (variable representation) or
individual (solution representation) over time.
– In general, higher mutation rate increases the versatility of the solutions within
the population, but it also increases the amount of randomness in the search.
• Mutation step size
– Also referred as σ .
– The size of the change of mutation rate compared to its previous value. In the
case of real-valued search spaces, mutation is usually performed by adding a
normally distributed random value to each optimised variable.
– In general, higher mutation step size has a similar influence as the higher
mutation rate.
• Crossover rate or crossover probability
– Also referred as p c .
– The probability that crossover is performed between two chosen solutions.
– In general, lower crossover probability prevents exchange of genetic material
and causes slower convergence.
• Selection pressure
– It is the degree to which the better individuals are favoured. Which solutions
will provide the individual with an increased chance of surviving over others,
i.e. to be forced to contribute to the next generation. Therefore, individuals
with certain phenotypes have an advantage when it comes to survival and
reproduction.
– The higher the selection pressure, the more good solutions are kept from
generation to generation. It is related to elitism strategy, which ensures good
solutions to proceed to the next generation.
• Tournament size
– Tournament selection involves running tournaments amongst pairs (or groups)
of individuals chosen at random from the population. The winner of each
tournament (the one with the best fitness) is selected for the next phase (i.e.
crossover).
– When tournament size is higher, the weaker solutions have a smaller chance
to be selected.
• Number of generations
– Also referred as n g .