摘要
An optimisation problem can have many forms and variants.It may consider different objectives,constraints,and variables.For that reason,providing a general application programming interface(API)to handle the problem data efficiently in all scenarios is impracticable.Nonetheless,on an R&D environment involving personnel from distinct backgrounds,having such an API can help with the development process because the team can focus on the research instead of implementations of data parsing,objective function calculation,and data structures.Also,some researchers might have a stronger background in programming than others,hence having a standard efficient API to handle the problem data improves reliability and productivity.This paper presents a design methodology to enable the development of efficient APIs to handle optimisation problems data based on a data-centric development framework.The proposed methodology involves the design of a data parser to handle the problem definition and data files and on a set of efficient data structures to hold the data in memory.Additionally,we bring three design patterns aimed to improve the performance of the API and techniques to improve the memory access by the user application.Also,we present the concepts of a Solution Builder that can manage solutions objects in memory better than built-in garbage collectors and provide an integrated objective function so that researchers can easily compare solutions from different solving techniques.Finally,we describe the positive results of employing a tailored API to a project involving the development of optimisation solutions for workforce scheduling and routing problems.
