drl-supply-chain-hilo-greedy.py

def find_optimal_price_t(p_baseline, price_grid, t):   # evalutes all possible price schedules 
  p_grid = np.tile(p_baseline, (len(price_grid), 1))   # derived from the baseline by  
  p_grid[:, t] = price_grid                            # changing the price at time t
  profit_grid = np.array([ profit_response(p) for p in p_grid ])
  return price_grid[ np.argmax(profit_grid) ]

p_opt = np.repeat(price_opt_const, T)                  # start with the constant price schedule
for t in range(T):                                     # and optimize one price at a time
  price_t = find_optimal_price_t(p_opt, price_grid, t)
  p_opt[t] = price_t

print(f'Achieved profit is {profit_response(p_opt)}')
plt.plot(range(len(p_opt)), p_opt, c='red')

#> Achieved profit is 2903637.2