A com­plex system can, in prin­ciple, be observ­able – that is, the system’s com­plete internal state can be recon­structed from its out­puts, which would osten­sibly involve describing in com­plete quan­ti­ta­tive detail all of its internal state vari­ables at once. In an actual exper­i­ment, how­ever, such mea­sure­ment is typ­i­cally beyond our reach, and so is lim­ited to a smaller number of those vari­ables. Referred to as sen­sors (or sensor nodes), these key vari­ables can be used to make the com­plete system observ­able. Recently, sci­en­tists at North­eastern Uni­ver­sity and MIT devised a graph­ical approach that first derives the math­e­mat­ical equa­tions describing a com­plex system’s dynamics, and then deter­mines the key sen­sors for that system. More­over, when applying their approach to bio­chem­ical reac­tion sys­tems, the researchers dis­cov­ered that the derived sen­sors were both nec­es­sary and suf­fi­cient to describe the com­plete system. The sci­en­tists con­clude that their find­ings allow a sys­tem­atic explo­ration of many diverse nat­ural, tech­no­log­ical and socioe­co­nomic systems.

Read the article at Phys.org →