Features request

[giacfalk]

Please comment here below with additional features request. Here is an initial tentative list:

• Full transition to terra (remove residual dependency from raster and rgdal)

• Currently, the package does not support facility-level attributes that can affect the location allocation or local density or size of the facility to be allocated (e.g. supply constraints such as beds per hospital or users per facilities), as all facilities are equally defined.

• The apporach to establish the location of the next facility to be allocated in the lattice space (in the continuous allocation problem) or the exact set of facilities optimizing the objective (in the discrete allocation problem) is currently based on heuristics which might not necessarily coincide with the single (if uniquely identified) globally optimal solution. The identification of a truly global optimal solution would however require the development of a constrained optimization framework requiring the use of computationally intensive professional solvers.

[giacfalk]

Thanks to Rafael Pereira for the following additional feature requests:

• allow one to use a custom travel time matrix as an input
• incorporate in the model service capacity (supply capacity of each facility. For example, the number of seats in a school or UCI beds in a hospital)

[giacfalk]

Another suggestion:

• Check the output of the allocation function in terms of demand fraction coverage when using weights and exponents in the definition of the demand. As this will be inconsistent with traveltime_stats function, which does not accept exponents and weights, yet (to be implemented)