Isolated dwarf galaxies are scientifically interesting because of several reasons: 1) According to the most acknowledged cosmological model, ΛCDM, galaxies form and grow by hierarchical merging of smaller galaxies into bigger ones. If this paradigm is correct, then the presently observed isolated dwarf galaxies represent primordial building blocks out of which massive galaxies formed; 2) Because of shallow gravity potentials of dwarf galaxies their evolution is very sensitive to various baryonic processes (stellar feedback, heating by intergalactic radiation field, etc.). This could be useful in constraining models of these processes; 3) Mass-luminosity ratios of dwarf galaxies are several orders of magnitude larger than in massive galaxies. This means that dwarf galaxies should have relatively a lot more dark matter and could be very important in determining its nature; 4) Isolated dwarf galaxies are simpler systems than massive galaxies (no AGN, no mergers etc.) so their evolution could be modeled more easily and more reliably.
In this work I will present a semi-analytic isolated dwarf galaxy evolution model which tracks the stochastic accretion of baryonic and dark matter, star formation, stochastic stellar feedback and chemical evolution of a galaxy. Its basic structure is similar to other semi-analytic galaxy formation and evolution models (e.g.[1, 2]). Stochastic growth of the dark matter halo is modeled by using merger trees generated by a Monte Carlo algorithm based on the extended Press-Schecter theory [3]. Star formation is modeled by using an analytic model of star formation rate in molecule-poor galaxies [4]. Stochastic stellar feedback and metal enrichment is modeled by generating random stellar masses from an empirical probability density function [5]. This model is used to investigate how stochastic galaxy growth and stellar feedback could affect the evolution of an isolated dwarf galaxy and its observable properties. We find that stochasticity can create significantly different star formation histories in galaxies with the same present dynamical masses and this results in up to 40% scatter in present-day stellar mass, gas mass and metallicity. Also we find that because of non-linear effects the most probable evolution and observable properties differ from the average model without stochastic effects. This might lead to disagreement between results of a model without stochastic effects and observations even if that model is correct.