Mathematical model for resource optimization of an enterprise information system based on performance and cost criteria

Abstract

The article addresses the relevant scientific and practical problem of developing a mathematical model for resource optimization of an enterprise information system based on performance and cost criteria. The relevance of the study is determined by the fact that modern enterprise information systems operate under dynamically changing workloads, limited computational, network, and financial resources, as well as increasing requirements for responsiveness, reliability, and service quality. Under such conditions, both insufficient and excessive resource provisioning negatively affect system efficiency. In the first case, overloads occur, response time increases, throughput decreases, and the share of SLA violations grows. In the second case, operational expenditures increase, which reduces the economic feasibility of maintaining the IT infrastructure.
The introduction substantiates the importance of enterprise information systems for ensuring business process continuity, supporting managerial decision-making, and maintaining organizationnal competitiveness. The problem statement highlights the need for a formalized approach to selecting a system configuration that would simultaneously consider technical and cost-related criteria. The literature review examines current approaches to resource optimization in enterprise, serviceoriented, cloud, and fog/edge environments, as well as studies devoted to the application of metaheuristic algorithms in resource allocation problems. It is established that despite the significant progress in this field, insufficient attention has been paid to the construction of a generalized mathematical model aimed specifically at optimizing the resources of an enterprise information system as an integrated object.
The purpose of the study is to develop a mathematical model for optimizing the resources of an enterprise information system that formalizes the process of selecting a rational configuration according to performance and cost criteria. Within the study, the enterprise information system is represented as a multi-component optimization object whose configuration is described by a vector of controllable parameters. The main evaluation indicators include average response time, throughput, utilization level of individual resources, the proportion of SLA violations, and the total cost of resource provisioning. On this basis, an objective function is proposed that integrates performance losses, configuration cost, and a penalty for service requirement violations. In addition, a system of constraints is formed to reflect the actual technological and economic conditions of system operation. The results of the study show that the proposed model can serve as a theoretical basis for further application of metaheuristic algorithms, particularly particle swarm optimization, differential evolution, and simulated annealing, in enterprise information system configuration problems. It is demonstrated that the multi-criteria nature of the problem and the complexity of the feasible solution space justify the use of such methods. The conclusions emphasize that the proposed mathematical model is sufficiently universal and adaptable to different types of enterprise information systems and can be extended toward simulation modeling, experimental validation, and hybrid algorithmic schemes for resource optimization.

Keywords: enterprise information system, resource optimization, mathematical model, performance, cost, metaheuristic algorithms, system configuration, SLA.