Genetic algorithm‐based price and warranty optimization in software systems

• Published in: Expert systems Vol. 41; no. 7
• Main Authors: Arora, Rajat, Tandon, Abhishek, Aggarwal, Anu G., Mittal, Rubina
• Format: Journal Article
• Language: English
• Published: Oxford Blackwell Publishing Ltd 01.07.2024
• Subjects: Applications programs; genetic algorithm; Genetic algorithms; Gumbel's bivariate distribution; Heuristic methods; Marketing; Optimization; Sales; Soft computing; Software development; Software engineering; Software industry; software reliability growth model; stochastic Bass model; two‐dimensional warranty; Warranties
• ISSN: 0266-4720; 1468-0394
• DOI: 10.1111/exsy.13334

Owing to increasing competition, the marketing of commercial software products is a challenging task. Diverse marketing strategies are adopted to tap the huge market share. Among these, pricing and warranty are the most significant ones. The warranty policies organizations adopt usually have two dimensions: duration and effort spent during the warranty phase. In this paper, two variables are broadly taken: testing, which is an essential phase during software development, and warranty, which is a crucial component of after‐sales service. Here, we will formulate an optimization problem where returns and expenses will be influenced by price, testing resources spent, the duration for which testing is done, warranty duration and effort spent during warranty for the software product. The product sales are assumed to follow a Stochastic Bass model based on Non‐Homogeneous Poisson Process (NHPP). A free on‐site support warranty with an upper limit on duration and effort covers the 2D failure process. The numerical illustration is presented wherein profit is maximized by conjointly optimizing five variables, that is price, testing duration, testing effort, warranty duration and warranty effort. This study emphasizes on application of a popular soft computing, metaheuristic technique named genetic algorithm (GA) to solve an optimization problem for a software system. The solution to the Optimization problem is obtained using the GA tool in MATLAB. The results obtained have relevance in the industry for finding optimal values of significant variables before the release of newly developed software products to the ultimate client. The concept discussed in this paper has diverse applications in computer engineering, software and hardware industry.