Differential Evolution-based Particle Swarm Optimization Algorithm (DEPSO)

The Differential Evolution-based Particle Swarm Optimization (DEPSO) algorithm is a hybrid optimization technique that integrates the advantages of both Differential Evolution (DE) and Particle Swarm Optimization (PSO). By incorporating DE's mutation and crossover operations, DEPSO enhances PSO's global search capability, effectively addressing traditional PSO's tendency to converge prematurely to local optima. In code implementation, this involves maintaining both velocity-position updates from PSO and differential mutation operations from DE within each iteration cycle.

In DEPSO implementation, each particle not only follows PSO's velocity update rules (using personal best and global best positions) but also generates new solutions through DE strategies. Specifically, at certain iteration intervals, the algorithm performs differential mutation on particles to create candidate solutions, followed by crossover operations to determine whether to replace current solutions. This mechanism enables particles to escape local optima while preserving PSO's fast convergence characteristics. The algorithmic flow typically includes parameter initialization, fitness evaluation, and conditional application of DE operations based on iteration counters or performance metrics.

DEPSO proves particularly effective for high-dimensional complex optimization problems such as engineering design optimization and parameter tuning scenarios, where DE's global exploration capability complements PSO's local exploitation strength. Experimental results demonstrate that this hybrid approach generally outperforms standalone PSO or DE algorithms in both convergence speed and solution accuracy. Key implementation considerations include balancing the frequency of DE operations, setting appropriate scaling factors for mutation, and optimizing crossover rates to maintain population diversity.