A MATHEMATICAL OPTIMIZATION MODEL FOR ADAPTIVE CACHING AND RESOURCE ALLOCATION IN DISTRIBUTED SYSTEMS

Abstract

The increasing adoption of cloud computing, edge computing, Internet of Things applications, content delivery services, distributed databases, mobile platforms, and data-intensive enterprise applications has created substantial challenges in the efficient utilization of distributed computing and communication resources. Large-scale distributed systems continuously process heterogeneous content requests while operating under limited cache capacity, computational resources, bandwidth, storage availability, energy constraints, and dynamically changing network conditions. Conventional caching and resource-allocation techniques frequently employ static placement policies or independently optimize storage, computing, and communication resources. Such approaches may perform poorly when user demand, content popularity, network congestion, node availability, and resource utilization change over time. This study proposes a Mathematical Optimization Model for Adaptive Caching and Resource Allocation in Distributed Systems (MOM-ACRA) that jointly determines content-placement decisions, cache utilization, computational-resource allocation, bandwidth assignment, and request-routing strategies. The proposed model formulates adaptive caching as a constrained multi-objective optimization problem designed to minimize content-access latency, communication cost, cache-miss rate, resource imbalance, and energy consumption while maximizing cache-hit probability, system throughput, resource utilization, and service availability. Binary content-placement variables are combined with continuous computing and communication allocation variables to represent decisions across heterogeneous distributed nodes.