Price optimization with machine learning

It would be preferred if the application is written in PHP since it is a part of CMS (Litecart), but other solutions that would include importing/exporting data would be accepted. Required functionality: 1) Changes buying, selling and trade-in price of products to the optimal value for maximum profits and also tries to get rid of excess stock 2) Two modes: automatic price adjustment every day at certain hour and the the applocation recommends the changes, but the user has to accept them 3) Must work on used and new products. New products can be reordered (lets say infitive supply), while used products are brought by customers, so the supply is limited and increasing buying price would probably get more products. 4) Suggest removing certain products from sales since cluttering the web page with products that do not sell can hurt overall sales 5) Suggest realocating stock between stores by looking at how well a certain products sells and at what store it is more likely to sell. Sometimes customers pick from a product range and in some cases the specific product is less important, as long as there are some options of this kind of product type. 6) Automatically adding and removing certain products as on sale price with limited time offer. 7) Display sales statistics, number of products sold, total price, revenue,.. per store and online sales. 8) Work with preorders, also adjust their price 9) Recommend stock for items that can be resupplied by entering the number of days the stock should last The following inputs are available, although some other could be added if possible: 1) quantity available (per store) 2) buying price for each product (note buying and tradein price are different) 3) date and selling price for each product sold 4) number of visitors for the product per day 5) how many times the product has been added to the basket per day (but not necessarly ordered) 6) how often the product has been ordered, but not actually bought (either the customer didn't pick up the product or rejected delivery) 7) in what store the product was purchased or if it was shipped to customer 8) type of product, for example sports, for children, racing 9) when the product will be available (preorders) For the current stock there is no info for what price the products were purchased, for the new products it will have to be added manually at the moment. The point of the application is that it is able to adjust to market quickly, there are big differences between time of the year, for example December has sales several times higher as during summer. Number of sales can often be more important as maximum profits, since it generates more sales in the long run.