Removed SP mock model.

README.md

@@ -35,23 +35,23 @@
 
 ### **Portfolio Optimization**
 
-For portfolio optimzation there are two paragimes which utilizes the simulation tensor. Stochastic Programming (SP) and Model Predictive Control (MPC). The main difference being SP utilizing all scenarios for one time period, while MPC considers the average scenario over multiple time periods. The implemented models yield different results and can be modified with additional constraints and asset classes to suit any investor.
+For portfolio optimzation there are two paradigmes which utilizes the simulation tensor. Stochastic Programming (SP) and Model Predictive Control (MPC). The main difference being SP utilizing all scenarios for one time period, while MPC considers the average scenario over multiple time periods. The implemented models yield different results and can be modified with additional constraints and asset classes to suit any investor.
 
 ##### 1. Maximize Expected Utility Domestic Stocks (Stochastic Programming)
 $$
-\mbox{max} \quad \mathbb{E}\Big[ U\Big( x_1^{\text{cash}} +  p_{1}^{\text{mid}} \boldsymbol{x}_{1}^{\text{hold}}  \Big)   \Big]
+\mbox{max} \quad \mathbb{E}\Big[ U\Big( x_1^{\text{cash}} +  \sum_{a \in A} p_{1,a}^{\text{mid}} x_{1,a}^{\text{hold}}  \Big)   \Big]
 $$
 
 $$
-\boldsymbol{x}_1^{\text{hold}} = \boldsymbol{h}_0^{\text{hold}} + \boldsymbol{x}_0^{\text{buy}} - \boldsymbol{x}_0^{\text{sell}}
+x_{1,a}^{\text{hold}} = h_{0,a}^{\text{hold}} + x_{0, a}^{\text{buy}} - x_{0,a}^{\text{sell}} \quad \forall a \in A
 $$
 
 $$
-x_1^{\text{cash}} = \big(h_0^{\text{cash}} +  p_0^{\text{mid}}(\boldsymbol{x}_0^{\text{buy}} - \boldsymbol{x}_0^{\text{sell}})   \big) e^{r \Delta t}
+x_1^{\text{cash}} = \big(h_0^{\text{cash}} +  \sum_{a \in A} (p_{0,a}^{\text{ask}}x_{0,a}^{\text{buy}} - p_{0,a}^{\text{bid}}x_{0,a}^{\text{sell}})   \big) e^{r \Delta t}
 $$
 
 $$
-\boldsymbol{x}_0^{\text{buy}}, \boldsymbol{x}_0^{\text{sell}}, \boldsymbol{x}_1^{\text{hold}} \geq \boldsymbol{0}
+x_{0, a}^{\text{buy}}, ~x_{0,a}^{\text{sell}}, ~x_{1, a}^{\text{hold}} \geq 0 \quad \forall a \in A
 $$
 
 $$

run.py

@@ -8,13 +8,13 @@
 from src.abacus.utils.portfolio import Portfolio
 from src.abacus.simulator.simulator import Simulator
 from src.abacus.assessor.risk_assessor import RiskAssessor
-from src.abacus.optimizer.optimizer import SPMaximumUtility, SPMaximumGain
+from src.abacus.optimizer.optimizer import SPMaximumUtility
 
 
 
 # Mock instrument creation...
 instruments = []
-for id, ticker in enumerate(sorted(TEST_YAHOO_STOCK_UNIVERSE_4)):
+for id, ticker in enumerate(sorted(TEST_YAHOO_STOCK_UNIVERSE_8)):
     file = f"tests/data/{ticker}.csv"
     time_series = pd.read_csv(file, index_col='Date')
     ins = Instrument(id, ticker, "Stock", time_series)
@@ -23,7 +23,7 @@
 # Simulation ...
 simulator = Simulator(instruments)
 simulator.calibrate()
-simulator.run_simulation(time_steps=10, number_of_simulations=1)
+simulator.run_simulation(time_steps=25, number_of_simulations=500)
 
 # Portfolio creation...
 holdings = 1, 1, 1, 1
@@ -36,23 +36,23 @@
 # assessor.summary()
 
 # Display reasonableness of simulations...
-# for i in range(25):
-#     y = simulator.price_tensor[0,:,i]
-#     x = [i for i in range(len(y))]
-#     pyplot.plot(x, y)
-# pyplot.show()
+for i in range(25):
+    y = simulator.price_tensor[0,:,i]
+    x = [i for i in range(len(y))]
+    pyplot.plot(x, y)
+pyplot.show()
 
 
 # Mock prices...
 price_tensor = torch.tensor([ [[1000]], [[0]], [[0]], [[0]]])
 inital_prices = torch.tensor([ [[10]], [[10]], [[10]], [[10]]])
 
 # Create optimizer with different optimization models...
-optimizer = SPMaximumGain(portfolio, price_tensor, inital_prices)
+optimizer = SPMaximumUtility(portfolio, simulator.price_tensor, simulator._inital_prices, gamma=1)
 optimizer.solve()
 
 print()
-optimizer = SPMaximumUtility(portfolio, price_tensor, inital_prices)
+optimizer = SPMaximumUtility(portfolio, simulator.price_tensor, simulator._inital_prices, gamma=-2)
 optimizer.solve()
 
 

src/abacus/optimizer/optimizer.py

@@ -54,50 +54,14 @@ def _check_initialization(self):
 
 
 
-class SPMaximumGain(OptimizationModel):
-
-    _model_specification = OptimizationSpecifications.SP_MAXIMIZE_GAIN
-
-    def __init__(self, portfolio: Portfolio, price_tensor: torch.Tensor, inital_prices: torch.Tensor):
-        super().__init__(portfolio, price_tensor)
-        self._inital_prices = inital_prices
-
-    def solve(self):
-        super().solve()
-        print(self._ampl.get_variable("x_buy").get_values())
-        print(self._ampl.get_variable("x_sell").get_values())
-        print(self._ampl.eval("display Objective;"))
-
-    def _set_ampl_data(self):
-        assets = self._portfolio.instruments
-        asset_identifiers = [instrument.identifier for instrument in assets]
-        instrument_holdings = np.array(list(self._portfolio.holdings.values()))
-        price_tensor = np.array(self._simulation_tensor[:,-1,:])
-        inital_prices =  dict(zip(asset_identifiers, np.array(self._inital_prices.reshape(4))))
-        tensor_size = price_tensor.shape
-        number_of_assets = tensor_size[0]
-        number_of_scenarios = tensor_size[1]
-        price_dict = {(j+1, asset.identifier): price_tensor[asset.id][j] for asset in assets for j in range(number_of_scenarios)}
-
-        self._ampl.get_set("assets").set_values(asset_identifiers)
-        self._ampl.param["risk_free_rate"] = 0.05
-        self._ampl.param["dt"] = 10/365
-        self._ampl.param["number_of_assets"] = number_of_assets
-        self._ampl.param["number_of_scenarios"] = number_of_scenarios
-        self._ampl.param["inital_cash"] = self._portfolio._cash
-        self._ampl.param["inital_holdings"] = instrument_holdings
-        self._ampl.param["inital_prices"] = inital_prices
-        self._ampl.param["prices"] = price_dict
-
-
-
 class SPMaximumUtility(OptimizationModel):
 
     _model_specification = OptimizationSpecifications.SP_MAXIMIZE_UTILITY
 
-    def __init__(self, portfolio: Portfolio, price_tensor: torch.Tensor, inital_prices: torch.Tensor):
+    def __init__(self, portfolio: Portfolio, price_tensor: torch.Tensor, inital_prices: torch.Tensor, gamma: float):
         super().__init__(portfolio, price_tensor)
         self._inital_prices = inital_prices
+        self._gamma = gamma
 
     def solve(self):
         super().solve()
@@ -110,7 +74,7 @@ def _set_ampl_data(self):
         asset_identifiers = [instrument.identifier for instrument in assets]
         instrument_holdings = np.array(list(self._portfolio.holdings.values()))
         price_tensor = np.array(self._simulation_tensor[:,-1,:])
-        inital_prices =  dict(zip(asset_identifiers, np.array(self._inital_prices.reshape(4))))
+        inital_prices =  dict(zip(asset_identifiers, np.array(self._inital_prices.reshape(8))))
         tensor_size = price_tensor.shape
         number_of_assets = tensor_size[0]
         number_of_scenarios = tensor_size[1]
@@ -119,7 +83,7 @@ def _set_ampl_data(self):
         self._ampl.get_set("assets").set_values(asset_identifiers)
         self._ampl.param["risk_free_rate"] = 0.05
         self._ampl.param["dt"] = 10/365
-        self._ampl.param["gamma"] = -2
+        self._ampl.param["gamma"] = self._gamma
         self._ampl.param["number_of_assets"] = number_of_assets
         self._ampl.param["number_of_scenarios"] = number_of_scenarios
         self._ampl.param["inital_cash"] = self._portfolio._cash

src/abacus/utils/enumerations.py

@@ -4,5 +4,4 @@
 
 class OptimizationSpecifications(Enum):
     SP_MAXIMIZE_UTILITY = "sp_maximize_utility.mod"
-    SP_MAXIMIZE_GAIN = "sp_maximize_gain.mod"
     MPC_MAXIMIZE_UTILITY = "mpc_maximize_utility.mod"