Utilitarian calculus has always been known to produce morally indefensible outcomes.
We use our moral sense to evaluate moral theories that match it, and Utilitarianism does pretty well in a lot of cases. But in some circumstances it is -- awful. The hospital example, where a patient walks in for a minor procedure, and the utilitarian doctors realize he just happens to be a perfect donor for 12 different critical patients who need organs ASAP, and they cut the unwitting donor up for a 12:1 improvement in lives saved that day! But ask even the families of the saved patients, and they would tell the docs not to kill another patient to save their loved one!
In other cases, the unpredictability of the future can be mismanaged/rationalized to justify all present evils. This was done in Soviet theology. The certainty of perpetual working-class misery without a Soviet revolution, and the benefits in perpetuity of a Workers Paradise, summed to infinity in a utilitarian calculus, and justified all the tortures, murders, and Gulags of the Soviet state. The efforts to run the calculus withOUT future humans, to avoid these sorts of rationalized evils, lead to a government policy that utterly fails to benefit subsequent generations.
So, the basic calculus methodology is suspect.
Additionally, your chosen value system -- a "hedonistic view of harms and benefits", is likewise well known to be seriously flawed. Under hedonism, one avoids any physical or mental stressors. But humans raised under hedonism, become useless. All skills we learn, are acquired by challenging ourselves, and overcoming those challenges. Hedonistically raised children become whiny and ineffective in all of life.
For actual humans, pleasure is a very transient "value", and "pain" likewise. What we actually need to optimize is harms and welfares. Teaching a child to love exercise, to love mental exploration and the development of new skills, while celebrating the overcoming of pains and frustrations in the process, is an essential task in raising them. Harms and welfares are much harder to "quantify" than pleasures and pains -- yet for utilitarianism to be at all useful in raising children, OR in planning one's own course thru life, this is the alternative calculus that one must make.
Many people have also argued that a calculation that is impossible in practice, as utilitarian calculus's are, is not an actually viable theory of moral decision making. This has led in recent years to a revival in Virtue Ethic thinking, primarily due to Virtue Ethic's focus on character development being so much more compatible with the welfare feature of harm/welfare.
Virtue ethics have their own failings, primarily around the "harm" side of that dichotomy, so switching to a purely virtue ethic model would not eliminate cases where "moral theory X prescribes clearly immoral situations".
We also know that rationalism cannot justify itself, and it is intrinsically flawed per Godel's incompleteness theorems. We also know from the psychological studies of those humans who can reason, but have no valid judgement, that it leads to absurdly inappropriate behavior.
So that a purely rationalist argument, using a suspect methodology, and suspect value system, ends up arguing for what you already think to be an invalid conclusion, is not actually a surprise, and there may not BE anything wrong with the authors' reasoning.
The philosophic involves learning to question the walls of the boxes one thinks within, and realizing that those walls are not certain. All mental boxes can limit one's ability to arrive at a better judgement about the world. The author's self-declared mental boxes in this paper are assumptions one must learn when to opt out from.