Optimisation

Description

Issues Addressed

Relationship with Other Components

Further Reading

---

To Do

• mathematical
• “Utility function” (and its other name: "objective function"?)
• SVG: see Barb\Peter's\Objective Function
• cf philosophy of utilitarianism qv
• It's not for me to lecture others on the goal they should adopt, but to point out that they have one and that it's probably based on their priorities.
• Put all GW stuff in read-less components.
• Possible goal: Minimise overall discrepancy between achievement and goal IAW priorities qv for all needs. Make human impact on the planet is as sustainable as possible. Maximise aggregate human happiness over time. Balance planet's resources between humans, other life, and other natural features. Balance rational necessity with ethical concerns'[BW]. Balance between priorities in general[BW], possibly including granite.
• Optimisation involves compromise: a willingness to trade off one priority qv for another. Sometimes this happens automatically; eg, a homeless person won't notice the absence of beautiful works of art around the doorway in which they sleep. However, some of us need to consciously decide between quality of life (steak) and GW.
• Should the balance between priorities qv change over time, as life and/or environment circumstances change?
• Why is life (esp human) so important? Nature created life and will destroy it. Why should we prolong the status quo? priorities qv. Is this a consequence of instinct? If so, should we trust it without conscious scrutiny?
• Should diversity be rewarded in the objective function? Should we aim for 100% humans and 0% other life forms, or not? If not, why not? This should arise from priorities qv.
• Examples of variables within objective function (what is the relationship between this and option-generation qv?):
  
  • Number of people.
  • Number of people moved off earth.
  • Extent of impact that each human has (which may be reduced by technology).
  • Quality of life.
  • Level of dependence on consumption as a contributor to quality of life.
  • Earth's ability to absorb human impact (resource consumption, pollution, GW, extinctions, etc).
• optimum is not nec ideal (perfect) because of constraints.
• Possible constraints (non-solutions): 'thou shalt not kill more than one billion people'; 'it can't cost more than $10 trillion'.
• Constraints: finite planetary resources (including the environment’s ability to absorb pollution) are relatively new constraints. Even if the objective function remains the same, the optimum solution may now be different (and have a lower value).
• The objective function should no longer be max(humans.count), despite what instinct qv keeps telling us.
• Sub-optimisation: We won't achieve global optimality by encouraging every individual to pursue their own happiness (emotion qv).
• non-solution; eg, killing 2 billion people: fails to meet a constraint (values) and/or couldn't possibly be implemented. cf no-good-options qv.

The axes in system-dynamics diagrams don't mean anything, whereas the past-now-future diagram at least has a time axis. Maybe what we need is a higher-dimensional system-dynamics diagram that has meaningful axes; eg, representing time and space (in additional to linkage between effects). There may be more dimensions too: probability? consequences? alternative decisions? Maybe we need a series of different views of the same model (inspired by MVM, UML).