Objectives: Widespread emergence of resistance to antimicrobial agents is a serious problem. The rate at which new agents are made available clinically is unlikely to keep up with these resistant pathogens, and there is an urgent need to accelerate antimicrobial agent development. We explored the use of mathematical modelling to guide selection of dosing regimens. Methods: Using time-kill studies data of Pseudomonas aeruginosa over 24 h, we developed a mathematical model to capture the dynamic relationship between a heterogeneous microbial population and meropenem concentrations. The microbial behaviour in response to meropenem over 5 days was predicted via computer simulation and subsequently validated using an in vitro hollow fibre infection model. Three parallel differential equations were used, each characterizing the rate of change of drug concentration, microbial susceptibility and microbial burden of the surviving population over time, respectively. Several model structures were explored; they differed in the adaptation of the microbial population over time. Various fluctuating concentration-time profiles of meropenem were experimentally examined, mimicking human elimination and repeated dosing. Results: Using limited experimental data as inputs, the mathematical model was reasonable in qualitatively predicting microbial response (sustained suppression or regrowth due to resistance emergence) to various pharmacokinetic profiles of meropenem. Conclusions: Our results suggest that mathematical modelling may be used to predict microbial response to a large number of antimicrobial agent dosing regimens efficiently, and have the potential to be used to guide highly targeted investigation of dosing regimens in pre-clinical studies and clinical trials. The in vivo relevance of the modelling approach warrants further investigations.