Where does swarm collide?

A cohesive group of animals that moves and reacts as a collective, such as a cloud of flying insects, a school of fish, or a flock of birds. The scientific study of why animals gather this way was founded by Allee's 1931 Animal Aggregations, the mathematics of how a swarm holds together and moves was set out in Okubo's 1986 Dynamical aspects of animal grouping, and Parrish and Edelstein-Keshet's 1999 Complexity, pattern, and evolutionary trade-offs in animal aggregation synthesised how such order arises. The referent is a living group whose pattern emerges from many individuals following simple local rules, and it is not a run of earthquakes, a search procedure on a computer, or a battlefield tactic.[1][2][3]

Unit

One aggregation of animals, sized by member count and packing density

Recovery means

Not applicable

Measured by

Quantified by group size, local density, alignment of individual headings, and correlation length between members

A burst of many earthquakes packed closely in time and place with no single dominant main shock, so the largest events are similar in size rather than one clear largest event followed by aftershocks. Mogi's 1963 Some discussions on aftershocks, foreshocks and earthquake swarms separated swarms from ordinary main shock and aftershock sequences and classified their patterns, and Hill's 1977 A model for earthquake swarms proposed how fluid moving through a mesh of small faults can drive one. The referent is a statistical pattern of ground rupture, often tied to volcanic or fluid activity, and it has nothing to do with grouped animals, an optimisation method, or a military manoeuvre.[4][5]

Unit

A time-and-space cluster of earthquakes lacking a dominant main shock

Recovery means

Not applicable

Measured by

Characterised by event count, magnitude distribution, spatial extent, and the duration of the cluster

A problem-solving method that turns loose many simple agents which share information and settle collectively on a good answer, imitating how living groups forage or move. The label swarm intelligence was introduced by Beni and Wang's 1989 work, published as Swarm intelligence in cellular robotic systems (1993), and its best known instance, particle swarm optimization, was set out in Kennedy and Eberhart's 1995 Particle swarm optimization, where candidate solutions drift through a search space pulled toward the best points found so far. The referent is a computational procedure run on a machine and judged by the quality of the answer it returns, and it is not a group of animals, a run of earthquakes, or a battlefield tactic.[6][7]

Unit

One population of candidate solutions iterating toward an optimum

Recovery means

Not applicable

Measured by

Assessed by solution quality, convergence speed, and the number of function evaluations used

A way of fighting in which many small, dispersed, networked units close on one target from several directions at once, strike together, then scatter and regroup. Arquilla and Ronfeldt's 2000 Swarming and the Future of Conflict named this pattern and argued it was becoming a dominant mode of conflict for networked forces, taking the label from the behaviour of insect and animal groups. The referent is a doctrine for arranging human units in space and time, studied in security and organisation research, and it is distinct from a biological aggregation, a seismic cluster, or a computer procedure.[8]

Unit

One coordinated convergence of dispersed units on a target

Recovery means

Not applicable

Measured by

Analysed by the number and dispersion of units, the timing of their convergence, and the communication among them