faty
Get Complete Project Material File(s) Now! »
Advantages of simulation modelling
Using simulation modelling to study evolution has many advantages. I have already alluded to some of the advantages, such as, the allowance of greater variation (especially individual variation) than analytical models (DeAngelis & Mooij, 2005; Grimm, 1999) and the natural inclusion of inclusive fitness has also been dealt with. I will mention a number of other advantages of simulation modelling, all of which, to some degree motivated its use, in this thesis: Individuals are discrete units in individual-based simulation. The system is thus investigated, using a ‘bottom-up’ approach where population characteristics emerge from the interactions of the individuals. This approach enables the investigator to track the behaviour of individual organisms (Grimm, 1999), which could aid in unravelling the system dynamics. When a researcher builds a simulation model, they usually tailor it to suit their research needs. In these models, we therefore have access to all the parameters, which we think are important in the system. Included are parameters that we cannot
manipulate in natural systems due to logistic, physiological, budget or ethical constraints (Grimm, 1999; Peck, 2004; Winsberg, 2003). With all the parameters under our control, it should be clear that an experimental approach may help a great deal in understanding the system. In spite of this, modellers often fail to perform methodical experiments (DeAngelis & Mooij, 2005; Grimm, 1999). Additional advantages of the customisability and flexibility of these models are that different data input formats can be handled, specific data output formats can be created (DeAngelis & Mooij, 2005) and it becomes easy to integrate empirical data into the models (Grimm, 1999).
Design concepts
Optimal mating strategies emerge from the population dynamics but the population dynamics are entirely characterized by rules specifying an individual’s behaviour. As mentioned, the fitness of each inbreeding class is specified explicitly and does not change during the simulation. The fitness of individuals more inbred than the last defined class have the same fitness as the individuals in last defined class. A fitness proportion scale with “roulette wheel” sampling is used to obtain the mating
individual (Mitchell, 1998). In brief, this means that the fitness of each individual is weighted, relative to the total fitness of the population, and assigned a proportion of the total fitness (i.e. fitter individuals have more numbers on the roulette wheel assigned to them). A random value is drawn between 0 and the total fitness of the population and the individual whose assigned proportion includes this value is selected to mate. Note that the fitness of haploid males was unaffected by their ancestry as all their loci are hemizygous and homozygosity is unaffected by their inbreeding history. Data gathered for analysis included the probability of inbreeding (α) in the population (calculated as the average inbreeding of the last 500 generations) as specified by the alleles of the individuals. It is important to note that inclusive fitness does not have to be introduced explicitly since kin-advantage will emerge by default in any individual-based simulation (Gros et al., 2008; Poethke et al., 2007).
1. INTRODUCTION
Individual-based simulation modelling
A simple simulation model
Stochasticity in simulation modelling
A short note on inclusive fitness in simulation modelling
Advantages of simulation modelling
Disadvantages of simulation modelling
Assessment of simulation models
Pollinating fig wasps
Aims of this thesis
2. EVOLUTIONARY STABLE MIXED MATING IN A VARIETY OF GENETIC SYSTEMS
Abstract
Introduction
Model description
State variables
Simulation process and scheduling
Design concepts
Initialisation and input
Sub-models
Results and Discussion
3. SEX RATIO DEPENDANT DISPERSAL WHEN SEX RATIOS VARY BETWEEN PATCHES
Abstract
Introduction
Model description
State variables and scales
Simulation process and scheduling (population life history)
Design Concepts
Initialization Input
Sub models
Statistics
Results and Discussion
4. ADEQUATE SAMPLE SIZES FOR ACCURATE DETECTION OF POPULATION SUBDIVISION: A SIMULATION BASED EXPLORATION OF SUMMARY STATISTICS
Abstract
Introduction
Materials and methods
Model description
General parameters
Experiments
Results
Discussion
Effect of population size on accuracy
Effect of the Fst on the accuracy
Effect of sample size and loci number on accuracy
Guidelines on sampling
5. INBREEDING DEPRESSION DOES NOT PROMOTE MIXED MATING AND DISPERSAL IN A MALE POLLINATING FIG WASP, PLATYSCAPA AWEKEI
Abstract
Introduction
Materials and Methods
Sample collection
Genotype reconstruction
Genotyping
Statistics
Results
Discussion
6. CONCLUSIONS
Mixed mating
Dispersal
Sampling
Platyscapa awekei
7. REFERENCES
