Particle Swarm Optimization Algorithms Inspired by Immunity-Clonal Mechanism and Their Applications to Spam Detection

Compared to conventional PSO algorithm, particle swarm optimization algorithms inspired by immunity-clonal strategies are presented for their rapid convergence, easy implementation and ability of optimization. A novel PSO algorithm, clonal particle swarm optimization (CPSO) algorithm, is proposed based on clonal principle in natural immune system. By cloning the best individual of successive generations, the CPSO enlarges the area near the promising candidate solution and accelerates the evolution of the swarm, leading to better optimization capability and faster convergence performance than conventional PSO. As a variant, an advance-and-retreat strategy is incorporated to find the nearby minima in an enlarged solution space for greatly accelerating the CPSO before the next clonal operation. A black hole model is also established for easy implementation and good performance. Detailed descriptions of the CPSO algorithm and its variants are elaborated. Extensive experiments on 15 benchmark test functions demonstrate that the proposed CPSO algorithms speedup the evolution procedure and improve the global optimization performance. Finally, an application of the proposed PSO algorithms to spam detection is provided in comparison with the other three methods.

Constraint Handling in Particle Swarm Optimization

In this article, the authors propose a particle swarm optimization (PSO) for constrained optimization. The proposed PSO adopts a multiobjective approach to constraint handling. Procedures to update the feasible and infeasible personal best are designed to encourage finding feasible regions and convergence toward the Pareto front. In addition, the infeasible nondominated solutions are stored in the global best archive to exploit the hidden information for guiding the particles toward feasible regions. Furthermore, the number of feasible personal best in the personal best memory and the scalar constraint violations of personal best and global best are used to adapt the acceleration constants in the PSO flight equations. The purpose is to find more feasible particles and search for better solutions during the process. The mutation procedure is applied to encourage global and fine-tune local searches. The simulation results indicate that the proposed constrained PSO is highly competitive, achieving promising performance.

Taguchi-Particle Swarm Optimization for Numerical Optimization

In this work, a hybrid Taguchi-Particle Swarm Optimization (TPSO) is proposed to solve global numerical optimization problems with continuous and discrete variables. This hybrid algorithm combines the well-known Particle Swarm Optimization Algorithm with the established Taguchi method, which has been an important tool for robust design. This paper presents the improvements obtained despite the simplicity of the hybridization process. The Taguchi method is run only once in every PSO iteration and therefore does not give significant impact in terms of computational cost. The method creates a more diversified population, which also contributes to the success of avoiding premature convergence. The proposed method is effectively applied to solve 13 benchmark problems. This study’s results show drastic improvements in comparison with the standard PSO algorithm involving continuous and discrete variables on high dimensional benchmark functions.

Networks Do Matter

This article describes a socially motivated evolutionary algorithm, Cultural Algorithms, to design a controller for a 3D racing game for use in a competitive event held at the 2008 IEEE World Congress. The controller was modeled as a state machine and a set of utility functions were associated with actions performed in each state. Cultural Algorithms are used to optimize these functions. Cultural Algorithms consist of a Population Space, a collection of knowledge sources in the Belief Space, and a communication protocol connecting the components together. The knowledge sources in the belief space vie to control individuals in the population through the social fabric influence function. Here the population is a network of chromosomes connected by the LBest topology. This LBest configuration was employed to train the system on an example oval track prior to the contest, but it did not generalize to other tracks. The authors investigated how other topologies performed when learning on each of the contest tracks. The square network (a type of small world network) worked best at distributing the influence of the knowledge sources, and reduced the likelihood of premature convergence for complex tracks.

Adaptive Neuro-Fuzzy Control Approach Based on Particle Swarm Optimization

This paper proposes a modified particle swarm optimization algorithm (MPSO) to design adaptive neuro-fuzzy controller parameters for controlling the behavior of non-linear dynamical systems. The modification of the proposed algorithm includes adding adaptive weights to the swarm optimization algorithm, which introduces a new update. The proposed MPSO algorithm uses a minimum velocity threshold to control the velocity of the particles, avoids clustering of the particles, and maintains the diversity of the population in the search space. The mechanism of MPSO has better potential to explore good solutions in new search spaces. The proposed MPSO algorithm is also used to tune and optimize the controller parameters like the scaling factors, the membership functions, and the rule base. To illustrate the adaptation process, the proposed neuro-fuzzy controller based on MPSO algorithm is applied successfully to control the behavior of both non-linear single machine power systems and non-linear inverted pendulum systems. Simulation results demonstrate that the adaptive neuro-fuzzy logic controller application based on MPSO can effectively and robustly enhance the damping of oscillations.

Beyond Standard Particle Swarm Optimisation

Currently, two very similar versions of PSO are available that could be called “standard”. While it is easy to merge them, their common drawbacks still remain. Therefore, in this paper, the author goes beyond simple merging by suggesting simple yet robust changes and solving a few well-known, common problems, while retaining the classical structure. The results can be proposed to the “swarmer community” as a new standard.

A Theoretical Framework for Estimating Swarm Success Probability Using Scouts

This paper presents a novel theoretical framework for swarms of agents. Before deploying a swarm for a task, it is advantageous to predict whether a desired percentage of the swarm will succeed. The authors present a framework that uses a small group of expendable “scout” agents to predict the success probability of the entire swarm, thereby preventing many agent losses. The scouts apply one of two formulas to predict – the standard Bernoulli trials formula or the new Bayesian formula. For experimental evaluation, the framework is applied to simulated agents navigating around obstacles to reach a goal location. Extensive experimental results compare the mean-squared error of the predictions of both formulas with ground truth, under varying circumstances. Results indicate the accuracy and robustness of the Bayesian approach. The framework also yields an intriguing result, namely, that both formulas usually predict better in the presence of (Lennard-Jones) inter-agent forces than when their independence assumptions hold.

Congestion Management Using Hybrid Particle Swarm Optimization Technique

This paper proposes the Hybrid Particle Swarm Optimization (HPSO) method for solving congestion management problems in a pool based electricity market. Congestion may occur due to lack of coordination between generation and transmission utilities or as a result of unexpected contingencies. In the proposed method, the control strategies to limit line loading to the security limits are by means of minimum adjustments in generations from the initial market clearing values. Embedding Evolutionary Programming (EP) technique in Particle Swarm Optimization (PSO) algorithm improves the global searching capability of PSO and also prevents the premature convergence in local minima. A number of functional operating constraints, such as branch flow limits and load bus voltage magnitude limits are included as penalties in the fitness function. Numerical results on three test systems namely modified IEEE 14 Bus, IEEE 30 Bus and IEEE 118 Bus systems are presented and the results are compared with PSO and EP approaches in order to demonstrate its performance.

Design of Multi-Criteria PI Controller Using Particle Swarm Optimization for Multiple UAVs Close Formation

Close formation flight is one of the most complicated problems on multiple Uninhabited Aerial Vehicles (UAVs) coordinated control. This paper proposes a new method to achieve close formation tracking control of multiple UAVs by applying Particle Swarm Optimization (PSO) based Proportional plus Integral (PI) controller. Due to its simple structure and effectiveness, multi-criteria PI control strategy is employed to design the controller for multiple UAVs formation, while PSO is used to optimize the controller parameters on-line. With the inclusion of overshoot, rise time, and system accumulated absolute error in the multi-criteria performance index, the overall performance of multi-criteria PI controller is optimized to be satisfactory. Simulation results show the feasibility and effectiveness of the proposed approach.

Distributed Multi-Agent Systems for a Collective Construction Task based on Virtual Swarm Intelligence

In this paper, a virtual swarm intelligence (VSI)-based algorithm is proposed to coordinate a distributed multi-robot system for a collective construction task. Three phases are involved in a construction task: search, detect, and carry. Initially, robots are randomly located within a bounded area and start random search for building blocks. Once the building blocks are detected, agents need to share the information with their local neighbors. A distributed virtual pheromone-trail (DVP) based model is proposed for local communication among agents. If multiple building blocks are detected in a local area, agents need to make decisions on which agent(s) should carry which block(s). To this end, a virtual particle swarm optimization (V-PSO)-based model is developed for multi-agent behavior coordination. Furthermore, a quorum sensing (QS)-based model is employed to balance the tradeoff between exploitation and exploration, so that an optimal overall performance can be achieved. Extensive simulation results on a collective construction task have demonstrated the efficiency and robustness of the proposed VSI-based framework.