Application of high-level methods of compromise optimization for control of autonomous robotized open pit mining

Abstract

In most software engineering approaches, software design begins with defining functional requirements, which is well suited to web-based software development projects. When designing high-critical large-scale software intended for industrial use, accounting for non-functional software requirements is also required. The main idea of the proposed document-oriented approach is to design a stable architectural solution as early as possible, taking into account the nonfunctional characteristics of the software: reliability, security, maintainability and performance (quality attributes). At the same time, the key issue is the coordination of functional requirements, taking into account technical limitations and business requirements achieved during the steady interaction of customer and developer teams. To increase the flexibility of the designed solutions and prevent crisis situations when developing highly critical large-scale software, it is proposed to use the approach integrating the architecture-centric design method (ACDM), the architecture-tradeoff analysis method (ATAM) with a matrix enterprise architecture matrix (EAM). This allows getting a result that is adequate to the required level of responsibility and reliability. Consideration of quality attributes within the framework of the method of compromise analysis makes it possible to select and make certain decisions in software design taking into account the scale of the software and its scope. The main attributes of product quality are highlighted (ISO 25010 standard), critical scenarios are defined for each of them (templates and use cases). The use of these templates for detailed software design with the necessary parameters of functional requirements, business conditions and technological limitations reduces the risk of developing unpredictable and uncertain system behavior. Based on the proposed approach, an architectural solution is presented for highly critical, responsible, large-scale software for managing autonomous robotic open-pit mining of minerals. Critical attributes for creating the specified software were identified and ranked, and the architecture of the solution according to the SWEBOK software development standard was described. Further, taking into account the nature, scale and scope of the software solution, recommendations are given on high-level architectural templates for the system design, including layers, pipelines and microservices. The proposed architecture-oriented development method is suitable for industrial-level software in various subject areas.