In TDC2 gateway data, we only have two light curve and have only one time delay parameter. However, we know in real word, we'll have multiple light curve (mostly 4), and have more time delay parameters. Therefore, to deal with real data, we need to be able to draw samples from multi dimensional time-delay parameter space.
In practice I think we only need to be able to cope with 1 or 3 parameters, for doubles and quads. But yes: the 3D problem is significantly harder than the 1D problem!
In TDC2 gateway data, we only have two light curve and have only one time delay parameter. However, we know in real word, we'll have multiple light curve (mostly 4), and have more time delay parameters. Therefore, to deal with real data, we need to be able to draw samples from multi dimensional time-delay parameter space.