The minimax regret method selects the alternative with the smallest worst-case regret when scenario probabilities are unknown or unreliable. Regret is the difference between an alternative’s payoff and the best payoff available in the same state of nature. section 10.0 section 10.1
Apply the method in four steps:
- Build a regret table by subtracting each payoff from the highest payoff in its scenario. section 10.1.3
- Find each alternative’s maximum regret across scenarios. section 10.1.3
- Choose the alternative with the smallest maximum regret. section 10.1.3
Investment alternatives under two economic states
For Strong and Weak Economy columns, the calculated regrets are respectively (0, 155000), (40000, 55000), (20000, 105000), and (65000, 0). Their maximum regrets are 155000, 55000, 105000, and 65000; therefore, the selected alternative has maximum regret 55000. section 10.1.3
>>> regret = payoff_table.copy() #1 >>> max_strong = regret['Strong Economy'].max() #2 >>> max_weak = regret['Weak Economy'].max() #3 >>> regret['Strong'] = \ >>> max_strong - regret['Strong Economy'] #4 >>> regret['Weak'] = \ >>> max_weak - regret['Weak Economy'] #5 >>> print(regret[['Strong', 'Weak']]) #6 Strong Weak 0 0 155000 1 40000 55000 2 20000 105000 3 65000 0 >>> regret['Maximum'] = \ >>> regret[['Strong', 'Weak']].max(axis = 1) #7 >>> print(regret[['Strong', 'Weak', 'Maximum']]) #8 Strong Weak Maximum 0 0 155000 155000 1 40000 55000 55000 2 20000 105000 105000 3 65000 0 65000 >>> min_max_regret = regret['Maximum'].min() #9 >>> print(min_max_regret) #10 55000
Follow the Python workflow for copying the payoff table, computing scenario regrets, aggregating row maxima, and selecting the minimum. section 10.1.3
