Q. What is Dynamic Programming and Decision Science?
Foundational Principle for this Conversation: To explore how the advanced analytical method known as Dynamic Programming can help one to strategize more effectively in the game of life, solve complex and extreme problems and make more effective decisions.
Definition: Decision Science (DS): An interdisciplinary mathematical science that deals with the application of advanced analytical methods through the use of technology to help organizations make better decisions.
Dynamic Programming: A method within the discipline of Decision Science for solving complex problems by
breaking them down into simpler sub-problems.
STUDENT: Can all problems be broken down into simpler sub-problems?
LEWIS: There is a point where enough information is available to solve a problem without breaking it down any further. Thus, with these problems dynamic programming would be overkill. Dynamic programming is specifically applicable to problems exhibiting the properties of overlapping sub-problems which are only slightly smaller and of optimal substructure. When dynamic programming is applicable, the method takes far less time than other less specific methods often called “naive methods” by mathematicians.
STUDENT: Please explore dynamic programming in greater depth.
LEWIS: The core idea behind dynamic programming is quite basic and simple. In general, to solve a given problem, one needs to solve different parts of the problem (sub-problems), then combine the solutions of these sub-problems and by doing so come to an overall solution.
STUDENT: What are some of the benefits of taking this approach to solving problems in this way?
LEWIS: Often as we break down a problem into sub-problems, we find that many of these sub-problems are identical. Dynamic programming allows one to solve each sub-problem only once, thus reducing the number of computations, reducing the risk of error, and maximizing potential at the lowest cost.
STUDENT: What is the importance of this in solving complex and extreme problems?
LEWIS: Complex and extreme problems often contain a number of recurring sub-problems which are exponentially large. Dynamic programming can be useful with these types of problems.
STUDENT: Are there sub-categories within dynamic programming as well?
LEWIS: Yes. There is Top-down dynamic programming which describes the storing of the results of certain calculations, which are later used again since the completed calculation is a sub-problem of a larger calculation. Then there is Bottom-up dynamic programming which involves formulating a complex calculation as a recursive series of simpler calculations.
STUDENT: Let’s go even deeper and apply all that we have discussed here to economics and classical game theory.
LEWIS: This brings us back to the concept of reinforced learning. Reinforcement learning is often used to explain how equilibrium may arise under what is called bounded rationality. Bounded rationality enables one to make the best decision and to create the most effective and efficient strategy even when the decision-maker has neither the time and space nor the ability to arrive at an optimal solution.
STUDENT: How would a person interested in living a simple life, “in the moment”, and not seeking to constantly optimize at all, benefit from any of what we have discussed concerning Decision science?
LEWIS: None of us live in isolation. It is not possible. So along with our inner wisdom, our intuition and our sense of tacit knowing we will still in daily life be presented with situations that will require a logical or at least rational decision just to survive. So even if one was not seeking to optimize at all, one would still need to make choices. The idea of bounded rationality is that individuals strive to be rational, having first greatly simplified the choices available. Thus, instead of choosing from every time or place, the decision-maker chooses between a small number of times and/or places. The result may be that decision-makers become what mathematicians call satisfiers; they accept a satisfactory solution which is good enough for their purposes rather than finding the optimum answer.
STUDENT: Where can I learn more about the origins of the theory of bounded rationality?
LEWIS: Explore the writing in the 1950s of H. A. Simon. On the web you can read more at:
STUDENT: Where can I learn more about Decision Making and strategizing more effectively?
LEWIS: See the Conversations in LHAGT on:
The Basics of Decision Science
Understanding Complex Problems
Understanding Extreme Problems
STUDENT: Do you have any concluding thought on this subject as it relates to LHAGT?
LEWIS: The importance of LHAGT is that it is concerned with developing latent human potentialities — regarded as our natural endowment as human beings, but rarely brought to fruition. With advanced technologies and tools and with higher levels of consciousness, inner growth and development, we have the opportunity to explore real possibilities for effective strategies to prosper in the game of life and by doing so serve others on a greater level.
Lewis Harrison is a futurist, speaker, NPR affiliated Talk Show Host and a professional copywriter. He can be reached at LewisCoaches@gmail.com
He is the administrator of a FB group on game theory and business solutions.
Lewis is an Internet marketing expert and copywriter and the author of the book “Gamification for Business”