Neurons receive distributed synaptic inputs over complex branched dendritic trees. These dendrites fundamentally allow for connectivity [1] and allow complex computations in space and time [2–4], but also impose metabolic costs in terms of maintenance and signal propagation [5]. Both the costs and benefits of dendritic trees are influenced by local diameters. The diameter profiles of dendrites are known to affect how much synaptic signals attenuate [6], how well mRNA or proteins can circulate [7], how quickly action potentials backpropagate [8], and even the strength of long-term synaptic plasticity [9]. A number of previous studies have used different methods to analyse how dendritic diameters could be optimised for single functions such as current transfer from synapse to soma [10,11] or protein trafficking across branch points [7]; however the techniques used in these studies are specific to a single function and cannot be easily generalised or systematically applied across entire dendritic trees. Here, we introduce a suite of fitting algorithms that allow local diameters to be optimised throughout entire dendritic trees for any arbitrary numeric cost function. Examples of such cost functions include maximising synaptic transfer to the soma, maximising the distribution of somatically generated proteins or mRNAs, maximising the temporal separation of fixed inputs, or allowing for locally co-operative or competitive synaptic plasticity. Further, multiple distinct cost functions can be combined and the trade-offs real neurons make between different, potentially competing, objectives can be examined [12]. In particular, we show that optimisation for current transfer gives a diameter relationship at branch points that describes real datasets well [13].