Several challenges accompany the deployment of solar photovoltaic (PV) technology in residential construction, such as determining an optimum size and layout design for on-site utilisation that conforms to local roof-sloping practice. Common solar PV installation practice tends to prefer the Equator-facing orientation due to its maximised energy aggregate regardless of household load patterns. Solar PV applications in high-latitude regions encounter other challenges, such as significant seasonal variations in daylight hours and in the sun's path. These challenges result in a PV mismatch: (a) in winter, minimal PV-generated energy and high energy demand (due to space and hot-water heating), and (b) in summer, PV over-generation and reduced energy demand. This study aims to provide a framework that identifies the best possible layout placement and size by applying the generalised reduced gradient nonlinear optimisation algorithm. In this regard, we monitor at one-minute intervals the energy performance of eleven single-family homes in Edmonton, Canada, some of which are net-zero energy homes and others of which are energy-efficient homes. Results show that in temperate climates at high latitudes in the Northern hemisphere, a south-west facing solar PV system installed at a tilt angle 10 above or below the local latitude can significantly improve the self-consumption compared to common installation practice.