High-Utility Itemset Mining (HUIM) is very important in finding combination of items which is very influential in revenue or profit in any of the different industries like retail, manufacturing and logistics. Another contribution of the paper is the proposal of an Enhanced Particle Swarm Optimization (PSO) model which addresses the shortcomings of the Standard PSO to address the issue of scalability, adaptability and complexity on real-life dataset. Despite the fact that the standard PSO algorithm incorporates the swarm intelligence in updating the particle positions based on the individual and groups experience, it tends to converge prematurely and fails in dynamic pricing and transaction variability in addition to high dimensional data. The Enhanced PSO deals with these issues by introducing the dynamic velocity and position boundaries that limit the particle motion to the realm of realistic operation, and normalization of the fitness function that prevents the domination by the extreme values that leads to the imbalanced optimization. In addition, improved cognitive and social parameters allow the algorithm to adapt dynamically, to allow it to escape local optima and converge globally. The Enhanced PSO was tried on a retail problem and demonstrated to scale better and be of practical significance, i.e., optimization of inventory and pricing strategies, with a normalized fitness of 0.8276 and raw fitness of 200 than the Standard PSO. It may be further developed in the future with the help of hybrid optimization that employs PSO and other optimization algorithms such as Genetic Algorithms or Differential Evolution to achieve even better performance in such areas as predictive analytics and artificial intelligence.