Particle Swarm Optimization is a population-based, stochastic optimization algorithm inspired by the collective behavior of social organisms such as bird flocks or fish schools. Developed in the mid-1990s by James Kennedy and Russell Eberhart, PSO is widely used to solve complex, multidimensional optimization problems where mathematical approaches may struggle.
In PSO, a set of potential solutions called particles “flies” through the problem’s search space. Each particle has a position that represents a potential solution and a velocity that defines how the position changes. Over successive iterations, particles update their velocity and position based on:
- Personal best position (pBest): the best solution a particle has found so far.
- Global best position (gBest): the best solution found by the entire swarm.
This iterative adjustment combines exploration, which searches new areas of the solution space, with exploitation, which refines known good solutions. This allows the swarm to converge toward an optimal or near-optimal solution without requiring gradient information or an explicit mathematical model of the problem.
Key features of PSO include:
- Simplicity: It requires minimal parameter tuning compared to other metaheuristic methods.
- Versatility: It can handle nonlinear, non-differentiable, and high-dimensional problems.
- Adaptability: Variants of PSO are used for both continuous and discrete optimization.
In the context of Project Portfolio Management (PPM), PSO can help optimize resource allocation, project selection, and scheduling by balancing competing objectives such as cost, risk, and strategic value under uncertainty.