• matematica applicata
• neuroscienze
• ottimizzazione combinatoria

Both in the neurological human vision system and in certain HEP particle detectors a strong
reduction of information happens at an early stage; following previous works of D. Benedetti, G. Punzi and
M. del Viva we state an abstract optimality principle which, taking into account storage and bandwidth
limits of the system, could explain the most efficient way for this data filtering process to happen. We
develop a mathematical model which formalizes this principle into a 0-1 combinatorial optimization
problem, reformulate it as a simple optimal path problem on a graph, and propose a simple albeit
inefficient algorithm to solve it; we prove the solution is not unique and the problem suffers from a form
of instableness in this regard, and propose a possible approximation which partially solves the issues with
the exact algorithm. Finally we compare, for various values of the model parameters, the behavior of the
exact solution with the computationally much simpler heuristic solution proposed by the original authors,
and we find there’s a range of values for which they behave very similarly.