Well-posedness and convexity in vector optimization