This paper conveys the concept and idea of reliability-based design optimization (RBDO) to design a robust and reliable pin-fin heat sink for operating in an uncertain environment. In doing so, we formulate an RBDO heat sink design problem and incorporate a recently developed cell evolution method (CEM) to solve the formulated RBDO problem that aims to minimize entropy generation rate under the simultaneous variations caused by ambient temperature and the velocity of the air approaching the heat sink. From finite element analysis of the RBDO designed heat sinks, we found that the increment of the target reliability tends to create a more symmetric and uniform temperature distribution around the center of the heat sink and thereby presents a greater ability to resist the environmental variations. Besides, extensive simulation results reveal that the RBDO designed heat sink is superior to conventional deterministic ones in providing a much more robust and reliable heat dissipation performance when faced with environmental uncertainties. Furthermore, we report the visible findings about how the reliable solution varies depending upon the reliability indices for pin-fin heat sinks with inline and staggered arrangements. (C) 2014 Elsevier Ltd. All rights reserved.