An enhanced connected banking system optimizer incorporating triple mechanism for solving global optimization problems

Smarter search for complex real‑world decisions

From planning airline timetables to tuning medical AI, many modern problems boil down to searching for the “best possible combination” among countless options. Exact mathematical methods often choke on this complexity. This paper introduces an improved computer search method, the Enhanced Connected Banking System Optimizer (ECBSO), that mimics how banks interact and trade information to hunt for better solutions faster and more reliably.

Why traditional methods hit a wall

Classical optimization techniques work well when problems are tidy: relationships are smooth, and the landscape of possibilities is relatively simple. But real applications typically involve many variables, messy constraints, and landscapes full of peaks and valleys where a search can get stuck in a merely “good” answer instead of the best one. Metaheuristic algorithms were invented to cope with this messiness. They borrow ideas from nature, physics, or human behavior—such as evolution, bird flocks, or classroom teaching—to wander intelligently through huge search spaces without needing perfect mathematical information.

Banking as a blueprint for problem solving

The earlier Connected Banking System Optimizer (CBSO) treated banks as search agents. Each “bank” represents a candidate solution, and transactions between banks model how solutions share information and improve over time. CBSO cycles through exploration (trying very different options) and exploitation (refining the best ones found so far). However, the original design had three key flaws: banks shared too little information across the whole network, the switch from exploring to exploiting was rigidly tied to time instead of actual progress, and the search relied too heavily on a single star performer, often causing the system to freeze around a subpar answer. These limitations became more severe as problems grew larger and more intricate.

Three new tricks that make the search sharper

ECBSO keeps the banking metaphor but adds three powerful mechanisms. First, a dominant group guidance strategy looks at the best-performing banks as a team instead of focusing on just one star. By capturing how their choices vary together, the algorithm generates new candidate solutions that follow the “collective wisdom” of this elite group, improving both coverage of the search space and the quality of promising leads. Second, a guided learning strategy continuously measures how much recent solutions have been moving. If the search is roaming too wildly, the algorithm nudges it toward careful refinement; if it is barely moving, ECBSO pushes it to explore new territory. Third, a hybrid elite strategy blends the original banking idea with another approach called an equilibrium optimizer. Instead of chasing one winner, ECBSO refines several strong candidates in parallel, which helps the system escape local traps and converge more steadily.

Putting the new method to the test

To see whether these changes really help, the authors tested ECBSO on a demanding international benchmark known as CEC 2017, which collects 29 artificial problems designed to stress optimization methods in many ways—simple, bumpy, hybrid, and highly tangled landscapes, each in several dimensions. ECBSO was compared with the original CBSO and eight leading competitors from different algorithm families. Across all test sizes, ECBSO consistently ranked first. It found better answers on more problems, did so more reliably from run to run, and showed smoother, more predictable progress curves. Detailed statistical checks confirmed that these gains were not due to luck. The authors then applied ECBSO to real engineering design tasks with constraints and again observed superior solution quality and robustness, although at the cost of somewhat longer run times because of its more involved calculations.

What this means for everyday technology

In plain terms, ECBSO is a more reliable “smart search” engine for very hard design and planning problems. By learning from a group of strong candidates, adjusting how boldly it explores based on recent behavior, and polishing several top options at once, it is better at avoiding dead ends and homing in on truly high-quality solutions. While it may not be ideal for ultra–time-critical tasks, its higher accuracy and stability make it a promising tool for offline decisions in areas like power systems, engineering design, scheduling, and machine learning, where finding a better solution can save substantial cost or improve safety.

Citation: Qian, D., Cai, X., Feng, L. et al. An enhanced connected banking system optimizer incorporating triple mechanism for solving global optimization problems. Sci Rep 16, 7747 (2026). https://doi.org/10.1038/s41598-026-36149-2

Keywords: metaheuristic optimization, banking-inspired algorithm, global optimization, engineering design, search algorithm