Table 1.

Techniques and algorithms used for steady-state and dynamic MFA and FBA.

Type	Advantages	Disadvantages	Description	Ref.
Static Optimisation approach	Simple implementation Suitable for GeM models  Fast	Provides a simple, not very detailed solution   Cannot predict metabolic shifts	Separates culture period into intervals of pseudo steady-state and performs an FBA optimisation for each of them	[131]
				[132]
Dynamic Optimisation Approach	Detailed representation of metabolism   Can describe metabolic shifts	Accurate parameter estimation in differential equations necessary    Need to avoid overfitting	Performs optimization over the whole period of interest with the use of differential equations to describe biomass and media concentrations	[132]
				[133]
				[54]
DMFA	Calculates intracellular fluxes	Requires extracellular metabolite concentrations thus, cannot be used in underdetermined systems	Uses a linear spline function to calculate intracellular fluxes	[55]
			Describes intracellular fluxes using linear changes of the fluxes though time	[134]
Multi-objective optimisation	Uses the duality theorem to achieve optimality	Numerical challenges arising from the DAE formulation	Uses logarithmic barrier functions on the constraints of the primal and dual problem. Converts them to a DAE with the dynamic balance equations for the substrates	[135]
	Deals with LP infeasibility that can be caused during time integration	Requires careful objective function setting to achieve realistic solution	DFBA using lexicographic optimisation to deal with the LP feasibility problem	[136]