Given their technological and market maturity, lithium-ion batteries are increasingly being considered and used in grid applications to provide a host of services such as frequency regulation, peak shaving, etc. Charging and discharging these batteries causes degradation in their performance. Lack of data on degradation processes combined with requirement of fast computation have led to over-simplified models of battery degradation. In this work, the recent experimental evidence that demonstrates that degradation in lithium-ion batteries is non-linearly dependent on the operating conditions is incorporated. Experimental aging data of a commercial battery have been used to develop a scheduling model applicable to the time constraints of a market model. A decomposition technique that enables the developed model to give near-optimal results for longer time horizons is also proposed.