Manuel Schmitt

Bernd Bassimir, Manuel Schmitt, Rolf Wanka

We study the variant of Particle Swarm Optimization (PSO) that applies random velocities in a dimension instead of the regular velocity update equations as soon as the so-called potential of the swarm falls below a certain bound in this dimension, arbitrarily set by the user. In this case, the swarm performs a forced move. In this paper, we are interested in how, by counting the forced moves, the swarm can decide for itself to stop its movement because it is improbable to find better solution candidates as it already has found. We formally prove that when the swarm is close to a (local) optimum, it behaves like a blind-searching cloud, and that the frequency of forced moves exceeds a certain, objective function-independent value. Based on this observation, we define stopping criteria and evaluate them experimentally showing that good solution candidates can be found much faster than applying other criteria.

Alexander Raß, Manuel Schmitt, Rolf Wanka

Particle Swarm Optimization (PSO) is a nature-inspired meta-heuristic for solving continuous optimization problems. In the literature, the potential of the particles of swarm has been used to show that slightly modified PSO guarantees convergence to local optima. Here we show that under specific circumstances the unmodified PSO, even with swarm parameters known (from the literature) to be good, almost surely does not yield convergence to a local optimum is provided. This undesirable phenomenon is called stagnation. For this purpose, the particles' potential in each dimension is analyzed mathematically. Additionally, some reasonable assumptions on the behavior if the particles' potential are made. Depending on the objective function and, interestingly, the number of particles, the potential in some dimensions may decrease much faster than in other dimensions. Therefore, these dimensions lose relevance, i.e., the contribution of their entries to the decisions about attractor updates becomes insignificant and, with positive probability, they never regain relevance. If Brownian Motion is assumed to be an approximation of the time-dependent drop of potential, practical, i.e., large values for this probability are calculated. Finally, on chosen multidimensional polynomials of degree two, experiments are provided showing that the required circumstances occur quite frequently. Furthermore, experiments are provided showing that even when the very simple sphere function is processed the described stagnation phenomenon occurs. Consequently, unmodified PSO does not converge to any local optimum of the chosen functions for tested parameter settings.

Vanessa Lange, Manuel Schmitt, Rolf Wanka

Particle Swarm Optimization (PSO) is a popular nature-inspired meta-heuristic for solving continuous optimization problems. Although this technique is widely used, the understanding of the mechanisms that make swarms so successful is still limited. We present the first substantial experimental investigation of the influence of the local attractor on the quality of exploration and exploitation. We compare in detail classical PSO with the social-only variant where local attractors are ignored. To measure the exploration capabilities, we determine how frequently both variants return results in the neighborhood of the global optimum. We measure the quality of exploitation by considering only function values from runs that reached a search point sufficiently close to the global optimum and then comparing in how many digits such values still deviate from the global minimum value. It turns out that the local attractor significantly improves the exploration, but sometimes reduces the quality of the exploitation. As a compromise, we propose and evaluate a hybrid PSO which switches off its local attractors at a certain point in time. The effects mentioned can also be observed by measuring the potential of the swarm.

Manuel Schmitt, Rolf Wanka

Particle swarm optimization (PSO) is a widely used nature-inspired meta-heuristic for solving continuous optimization problems. However, when running the PSO algorithm, one encounters the phenomenon of so-called stagnation, that means in our context, the whole swarm starts to converge to a solution that is not (even a local) optimum. The goal of this work is to point out possible reasons why the swarm stagnates at these non-optimal points. To achieve our results, we use the newly defined potential of a swarm. The total potential has a portion for every dimension of the search space, and it drops when the swarm approaches the point of convergence. As it turns out experimentally, the swarm is very likely to come sometimes into "unbalanced" states, i. e., almost all potential belongs to one axis. Therefore, the swarm becomes blind for improvements still possible in any other direction. Finally, we show how in the light of the potential and these observations, a slightly adapted PSO rebalances the potential and therefore increases the quality of the solution.