protocol

particle_swarm_optimization_protocol

Protocol for continuous bounded particle swarm optimization problem definitions. A problem object must define the three required predicates and may optionally define predicates for initial velocities, stopping, and progress reporting.

Availability:

logtalk_load(particle_swarm_optimization(loader))

Author: Paulo Moura

Version: 1:0:0

Date: 2026-08-05

Compilation flags:

static

Dependencies:

(none)

Remarks:

(none)

Inherited public predicates:

(none)

Public predicates

initial_positions/1

Returns a non-empty list of initial particle positions. All positions must be non-empty lists of numbers with the same dimensions as the position bounds.

Compilation flags:

static

Template:

initial_positions(Positions)

Mode and number of proofs:

initial_positions(-list(list(number))) - one

---

position_bounds/1

Returns one Lower-Upper numeric bound pair per position dimension, with Lower =< Upper.

Compilation flags:

static

Template:

position_bounds(Bounds)

Mode and number of proofs:

position_bounds(-list(pair)) - one

---

initial_velocities/1

Returns one initial velocity vector per initial position. Each vector must have the same dimensions as the position bounds, with every component between plus or minus the corresponding position range. Optional. When not defined or when it fails, velocities are initialized randomly.

Compilation flags:

static

Template:

initial_velocities(Velocities)

Mode and number of proofs:

initial_velocities(-list(list(number))) - zero_or_one

---

fitness/2

Computes the fitness of a particle position. The optimization direction is selected using the optimizer objective/1 option.

Compilation flags:

static

Template:

fitness(Position,Fitness)

Mode and number of proofs:

fitness(+list(number),-number) - one

---

stop_condition/3

True when the search should stop given the completed iteration count, best position, and best fitness. Optional. When not defined, the search runs for the configured maximum number of iterations.

Compilation flags:

static

Template:

stop_condition(Iteration,BestPosition,BestFitness)

Mode and number of proofs:

stop_condition(+non_negative_integer,+list(number),+number) - zero_or_one

---

progress/5

Called periodically to report the completed iteration count, best position, best fitness, mean swarm fitness, and swarm diversity. Diversity is the mean Euclidean distance from the swarm centroid. Optional.

Compilation flags:

static

Template:

progress(Iteration,BestPosition,BestFitness,MeanFitness,Diversity)

Mode and number of proofs:

progress(+non_negative_integer,+list(number),+number,+number,+number) - zero_or_one

---

Protected predicates

(none)

Private predicates

(none)

Operators

(none)

See also

particle_swarm_optimization(Problem)