NewsVendor purchase cost error

[joefarrington]

Hi team,

It looks to me like there's an error in the _STEP method of NewsVendorEnv when calculating the purchase cost.

The line is

purchase_cost = excess_inventory * self.cost * order_qty * \
            self.gamma ** self.lead_time

But I don't know why excess_inventory would be needed to calculate the purchase cost.

For reference, the corresponding line from the original implementation by Balaji et al is appears to be the same except for the multiplication by excess inventory

purchase_cost = self.gamma ** self.l * c * buys

(see https://github.com/awslabs/or-rl-benchmarks/blob/master/News%20Vendor/src/news_vendor_environment.py, their self.l is the lead time, c is the purchase cost and buys is the order quantity).

[joefarrington]

Sorry, thought I'd searched properly but just seen this is a duplicate of #26