A multi-factor intelligent biologic search algorithm for closed-area fiber optic network planning

Abstract

Fiber optic network planning is a critical aspect in numerous intelligent park construction projects. This study addresses the issue of fiber routing allocation within these enclosed parks, which necessitate substantial real-time communication during their operation. To optimize route allocation efficiency and account for line costs, we propose a Multi-factor Intelligent Biologic Search Algorithm (MIBSA). This algorithm operates within a hierarchical framework and progresses through three stages of enhancement. The foundational model utilizes an enhanced Lifelong Planning A* (LPA*) algorithm to delineate the initial path and non-uniformly allocate the primary pheromone of the ant colony algorithm. The upper tier integrates an improved Ant Colony Optimization (ACO) model, incorporating various factors into the pheromone update strategy for multi-objective fiber optic network planning. Simultaneously, this algorithm introduces a batch pheromone boosting mechanism and a momentum method to expedite the convergence process. Experimental findings highlight that MIBSA surpasses the conventional ACO method in solving fiber optic network planning problems, yielding an average comprehensive cost reduction of 16.42% and 7.82% under the environmental map of the objective problem and the general environmental map test, respectively.