Algorithm portfolio
Family of AI algorithms developed through a sustained programme of research into heterogeneity, topology, cooperation, bio-inspired information exchange and ensemble optimisation.
A heterogeneous particle-swarm architecture that combines complementary search behaviours with adaptive neighbourhood structure.
Foundation & adaptive structureA hybrid optimiser using genetic operators to reinforce exploration and preserve useful diversity during global search.
Foundation & adaptive structureConditional energy lending and borrowing between particles, inspired by altruistic behaviour in biological populations.
Bio-inspired heterogeneous behaviourA search mechanism inspired by biological eavesdropping, allowing particles to exploit information generated by distinct behavioural groups.
Bio-inspired heterogeneous behaviourDynamic group formation and collective quorum mechanisms for balancing local refinement with global exploration.
Collective dynamics & new search designA distinct swarm-based optimisation method designed to explore complex continuous search landscapes.
Collective dynamics & new search designA purposefully engineered sequential ensemble in which AHPSO preserves diversity and discovers promising regions, HPSO-TVAC refines inherited solutions, and MaPSO drives late-stage exploitation and convergence. Tested against 39 algorithms across CEC'13/14/17/20, it ranked first on the 100-dimensional CEC'14 and CEC'17 suites and required no problem-specific parameter tuning.
Ensemble & general-purpose optimisationResearch lineage
Each method explores a distinct mechanism while contributing to a broader question: how can heterogeneous agents exchange information, resources and behavioural roles to improve the reliability of global optimisation?
Portfolio architecture
Complementary particle roles and structured multi-swarm units create multiple modes of learning within one population.
Altruism, eavesdropping, fission–fusion and quorum mechanisms become operational rules for collective search.
Topology adaptation, hybridisation and staged optimisation coordinate exploration, refinement and exploitation.