Software products have become an integral part of human lives, and therefore need to account for human values such as privacy, fairness, and equality. Ignoring human values in software development leads to biases and violations of human values: racial biases in recidivism assessment and facial recognition software are well-known examples of such issues. One of the most critical steps in software development is Software Release Planning (SRP), where decisions are made about the presence or absence of the requirements (features) in the software. Such decisions are primarily guided by the economic value of the requirements, ignoring their impacts on a broader range of human values. That may result in ignoring (selecting) requirements that positively (negatively) impact human values, increasing the risk of value breaches in the software. To address this, we have proposed an Integer Programming approach to considering human values in software release planning. In this regard, an Integer Linear Programming (ILP) model has been proposed, that explicitly accounts for human values in finding an "optimal" subset of the requirements. The ILP model exploits the algebraic structure of fuzzy graphs to capture dependencies and conflicts among the values of the requirements.
翻译:软件产品已成为人类生活的一个组成部分,因此需要考虑到隐私、公正和平等等人类价值观,软件开发中无视人类价值观会导致偏见和违反人类价值观:累犯评估和面部识别软件中的种族偏见是这类问题众所周知的例子。软件开发中最重要的步骤之一是软件释放规划(SRP),在软件释放规划(SSRP)中就软件要求(性能)的存在与否作出决定,这些决定主要以需求的经济价值为指南,忽视其对更广泛的人类价值观的影响。这可能导致忽视(选择)积极(消极)影响人类价值观的要求,增加软件价值破坏的风险。为解决这一问题,我们提出了在软件释放规划中考虑人类价值观的Integer规划方法。在这方面,提出了Integer线性规划模式,明确说明人类价值观在寻找需求中的“最佳”部分。ILP模型利用了模糊图表的代数结构来测定可靠性和冲突。