Calculating Akaike Weights for best fitting models

[Manuel Riveros]

We're testing some models for a population which went through recent and short bottlenecks.

Our best fitting model, using AIC, is the one with a constant population size and the second best model is the one which matches the historical data.

We want to know if it's possible to calculate Akaike weights similar to what DIYABC does, to see how more likely the first model is compared to the second in a cumulative scale ranging from 0-1.

Is that possible or even correct to do?

Thanks,

Manuel Escalona