Merge branch 'next_release' into master

.gitignore

@@ -1,7 +1,7 @@
-*pyc
+!.gitignore
+.*
+*.pyc
 **parse_table_datacmds.py
-.idea
-.DS_Store
 gurobi.log
 switch_model.egg-info/
 venv

CHANGELOG.md

@@ -1,14 +1,29 @@
+-------------------------------------------------------------------------------
+Switch 2.0.7
+-------------------------------------------------------------------------------
+Note that as of version 5.7, Pyomo no longer supports "inf", "infinity" or "nan"
+values in the input files, so Switch will likely report errors if you have
+those.
+
+Changed Pyomo dependencies to versions 5.5.1-6.4.2. Note that Pyomo versions
+before 5.7.3 don't work with CPLEX version 12.10 or later (see
+https://github.com/Pyomo/pyomo/pull/1792). If using a recent version of CPLEX,
+you should ensure that you also use version 5.7.3 or later of Pyomo.
+
+Renamed outputs/dispatch-wide.csv to outputs/dispatch_wide.csv
+
 -------------------------------------------------------------------------------
 Switch 2.0.6
 -------------------------------------------------------------------------------
-This release fixes a bug where the fixed costs from all possible build years
-were applied each period, instead of only using the build years that are still
-in service in the current period. This increased the apparent cost of storage by
-approximately (study length) / (storage life). This bug was introduced in 
-version 2.0.0b3 and persisted through version 2.0.5, so results from earlier
-models will need to be updated.
+This release fixes a bug where the fixed costs of storage energy capacity (the
+MWh part of storage) from all possible build years were mistakenly applied each
+period, instead of only using the build years that are still in service in the
+current period. This increased the apparent cost of storage by approximately
+(study length) / (storage life). This bug was introduced in version 2.0.0b3 and
+persisted through version 2.0.5, so results from earlier models will need to be
+updated.
 
-This will be the last version of Switch to work in Python 2. It requires at 
+This will be the last version of Switch to work in Python 2. It requires at
 least Python 2.7.12 and also works with Python 3.
 
 -------------------------------------------------------------------------------

INSTALL.md

@@ -108,7 +108,7 @@ that they also use to solve models. This supports a number of activities:
 
 To work this way, first install Switch as described above (this will install all
 the Switch dependencies, even though you will later reinstall Switch itself).
-Then, in a terminal window or Anaconda command prompt Anaconda command prompt,
+Then, in a terminal window or Anaconda command prompt,
 use the `cd` and `mkdir` commands to create and/or enter the directory where you
 would like to store the Switch model code and examples. Once you are in that
 directory, run the following commands (don't type the comments that start with

LICENSE

@@ -1,4 +1,6 @@
-The Pyomo implementation of the SWITCH model for planning least-cost low-emission electricity grids is licensed under Apache License v2.0 which can be obtained at http://www.apache.org/licenses/LICENSE-2.0
+The Switch model for planning least-cost low-emission electricity grids is
+licensed under Apache License v2.0 which can be obtained at
+http://www.apache.org/licenses/LICENSE-2.0.
 
 
                                  Apache License
@@ -201,4 +203,4 @@ The Pyomo implementation of the SWITCH model for planning least-cost low-emissio
    distributed under the License is distributed on an "AS IS" BASIS,
    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    See the License for the specific language governing permissions and
-   limitations under the License.
+   limitations under the License.

examples/3zone_toy/inputs/fuel_supply_curves.csv

@@ -1,10 +1,10 @@
 regional_fuel_market,period,tier,unit_cost,max_avail_at_cost
-All_DistOil,2020,0,21.9802,inf
+All_DistOil,2020,0,21.9802,.
 All_DistOil,2030,0,24.5216,100000.0
 All_NG,2020,0,4.4647,1950514555.0
-All_NG,2020,1,5.0709,inf
+All_NG,2020,1,5.0709,.
 All_NG,2030,0,5.925,2368354558.0
-All_NG,2030,1,5.925,inf
+All_NG,2030,1,5.925,.
 North_Bio,2020,0,1.7102,6864985.0
 North_Bio,2020,1,3.3941,6782413.0
 North_Bio,2030,0,2.0438,6064415.0

examples/3zone_toy/inputs/gen_build_predetermined.csv

@@ -1,4 +1,4 @@
-GENERATION_PROJECT,build_year,gen_predetermined_cap
+GENERATION_PROJECT,build_year,build_gen_predetermined
 N-Coal_ST,1995,2
 N-Geothermal,2000,1
 N-NG_CC,2008,2

examples/3zone_toy/inputs/load_zones.csv

@@ -1,4 +1,4 @@
-LOAD_ZONE,dbid,existing_local_td,local_td_annual_cost_per_mw
-North,1,5.5,66406.5
-Central,2,3.5,61663.4
-South,3,9.5,128040.0
+LOAD_ZONE,dbid,existing_local_td,local_td_annual_cost_per_mw,local_td_loss_rate
+North,1,5.5,66406.5,0.053
+Central,2,3.5,61663.4,0.053
+South,3,9.5,128040.0,0.053

examples/3zone_toy/inputs/switch_inputs_version.txt

@@ -1 +1 @@
-2.0.6
+2.0.7

examples/3zone_toy/inputs/trans_params.csv

@@ -1,2 +1,2 @@
-trans_capital_cost_per_mw_km,trans_lifetime_yrs,trans_fixed_om_fraction,distribution_loss_rate
-1000.0,20,0.03,0.053
+trans_capital_cost_per_mw_km,trans_lifetime_yrs,trans_fixed_om_fraction
+1000.0,20,0.03

examples/3zone_toy/outputs/total_cost.txt

@@ -1 +1 @@
-134733088.429
+134733088.42929107

examples/3zone_toy_stochastic_PySP/PySPInputGenerator.py

@@ -48,6 +48,7 @@
 
 """
 from __future__ import print_function
+
 # Inputs directory relative to the location of this script.
 inputs_dir = "inputs"
 # ScenarioStructure.dat and RootNode.dat will be saved to a
@@ -58,11 +59,10 @@
 stage_list = ["Investment", "Operation"]
 stage_vars = {
     "Investment": ["BuildGen", "BuildLocalTD", "BuildTx"],
-    "Operation": ["DispatchGen", "GenFuelUseRate"]
+    "Operation": ["DispatchGen", "GenFuelUseRate"],
 }
 # List of scenario names
-scenario_list = [
-    "LowFuelCosts", "MediumFuelCosts", "HighFuelCosts"]
+scenario_list = ["LowFuelCosts", "MediumFuelCosts", "HighFuelCosts"]
 
 ###########################################################
 
@@ -82,6 +82,7 @@
 instance = model.load_inputs(inputs_dir=inputs_dir)
 print("inputs successfully loaded...")
 
+
 def save_dat_files():
 
     if not os.path.exists(os.path.join(inputs_dir, pysp_subdir)):
@@ -92,8 +93,9 @@ def save_dat_files():
 
     dat_file = os.path.join(inputs_dir, pysp_subdir, "RootNode.dat")
     print("creating and saving {}...".format(dat_file))
-    utilities.save_inputs_as_dat(model, instance, save_path=dat_file,
-        sorted_output=model.options.sorted_output)
+    utilities.save_inputs_as_dat(
+        model, instance, save_path=dat_file, sorted_output=model.options.sorted_output
+    )
 
     #######################
     # ScenarioStructure.dat
@@ -117,7 +119,7 @@ def save_dat_files():
 
         f.write("param NodeStage := RootNode {}\n".format(stage_list[0]))
         for s in scenario_list:
-            f.write("    {scen} {st}\n".format(scen=s,st=stage_list[1]))
+            f.write("    {scen} {st}\n".format(scen=s, st=stage_list[1]))
         f.write(";\n\n")
 
         f.write("set Children[RootNode] := ")
@@ -127,7 +129,7 @@ def save_dat_files():
 
         f.write("param ConditionalProbability := RootNode 1.0")
         # All scenarios have the same probability in this example
-        probs = [1.0/len(scenario_list)] * (len(scenario_list) - 1)
+        probs = [1.0 / len(scenario_list)] * (len(scenario_list) - 1)
         # The remaining probability is lumped in the last scenario to avoid rounding issues
         probs.append(1.0 - sum(probs))
         for (s, p) in zip(scenario_list, probs):
@@ -150,14 +152,16 @@ def write_var_name(f, cname):
             if hasattr(instance, cname):
                 dimen = getattr(instance, cname).index_set().dimen
                 if dimen == 0:
-                   f.write("    {cn}\n".format(cn=cname))
+                    f.write("    {cn}\n".format(cn=cname))
                 else:
-                    indexing = (",".join(["*"]*dimen))
+                    indexing = ",".join(["*"] * dimen)
                     f.write("    {cn}[{dim}]\n".format(cn=cname, dim=indexing))
             else:
                 raise ValueError(
-                    "Variable '{}' is not a component of the model. Did you make a typo?".
-                    format(cname))
+                    "Variable '{}' is not a component of the model. Did you make a typo?".format(
+                        cname
+                    )
+                )
 
         for st in stage_list:
             f.write("set StageVariables[{}] := \n".format(st))
@@ -171,8 +175,9 @@ def write_var_name(f, cname):
         f.write("    Operation OperationCost\n")
         f.write(";")
 
+
 ####################
 
-if __name__ == '__main__':
+if __name__ == "__main__":
     # If the script is executed on the command line, then the .dat files are created.
     save_dat_files()

examples/3zone_toy_stochastic_PySP/ReferenceModel.py

@@ -37,7 +37,7 @@
 
 # Ideally, we would use the main codebase to generate the model, but the
 # mandatory switch argument parser is interferring with pysp's command line tools
-#model = switch_model.solve.main(return_model=True)
+# model = switch_model.solve.main(return_model=True)
 
 module_list = switch_model.solve.get_module_list(args=None)
 model = utilities.create_model(module_list, args=[])
@@ -53,14 +53,19 @@
 # are nested inside another function in the financials module, they can't
 # be called from this script.
 
+
 def calc_tp_costs_in_period(m, t):
-	return sum(
-		getattr(m, tp_cost)[t] * m.tp_weight_in_year[t]
-		for tp_cost in m.Cost_Components_Per_TP)
+    return sum(
+        getattr(m, tp_cost)[t] * m.tp_weight_in_year[t]
+        for tp_cost in m.Cost_Components_Per_TP
+    )
+
+
 def calc_annual_costs_in_period(m, p):
-	return sum(
-		getattr(m, annual_cost)[p]
-		for annual_cost in m.Cost_Components_Per_Period)
+    return sum(
+        getattr(m, annual_cost)[p] for annual_cost in m.Cost_Components_Per_Period
+    )
+
 
 # In the current version of Switch, all annual costs are defined
 # by First Stage decision variables, such as fixed O&M and capital
@@ -73,14 +78,19 @@ def calc_annual_costs_in_period(m, p):
 # decisions in this example.
 # Further comments on this are written in the Readme file.
 
-model.InvestmentCost = Expression(rule=lambda m: sum(
-                calc_annual_costs_in_period(m, p) * m.bring_annual_costs_to_base_year[p]
-                for p in m.PERIODS))
-
-model.OperationCost = Expression(rule=lambda m:
-	sum(
-		sum(calc_tp_costs_in_period(m, t) for t in m.TPS_IN_PERIOD[p]
-		   ) * m.bring_annual_costs_to_base_year[p]
-	for p in m.PERIODS))
+model.InvestmentCost = Expression(
+    rule=lambda m: sum(
+        calc_annual_costs_in_period(m, p) * m.bring_annual_costs_to_base_year[p]
+        for p in m.PERIODS
+    )
+)
+
+model.OperationCost = Expression(
+    rule=lambda m: sum(
+        sum(calc_tp_costs_in_period(m, t) for t in m.TPS_IN_PERIOD[p])
+        * m.bring_annual_costs_to_base_year[p]
+        for p in m.PERIODS
+    )
+)
 
 print("model successfully loaded...")

examples/3zone_toy_stochastic_PySP/inputs/gen_build_predetermined.csv

@@ -1,4 +1,4 @@
-GENERATION_PROJECT,build_year,gen_predetermined_cap
+GENERATION_PROJECT,build_year,build_gen_predetermined
 N-Coal_ST,1995,2
 N-Geothermal,2000,1
 N-NG_CC,2008,2

examples/3zone_toy_stochastic_PySP/inputs/switch_inputs_version.txt

@@ -1 +1 @@
-2.0.6
+2.0.7

examples/3zone_toy_stochastic_PySP/pha_bounds_cfg.py

@@ -4,20 +4,24 @@
 # Use this by adding terms like the following to the runph command:
 # --linearize-nonbinary-penalty-terms=5 --bounds-cfgfile=pha_bounds_cfg.py
 
+
 def pysp_boundsetter_callback(self, scenario_tree, scenario):
-    m = scenario._instance 	# see pyomo/pysp/scenariotree/tree_structure.py
+    m = scenario._instance  # see pyomo/pysp/scenariotree/tree_structure.py
 
     # BuildLocalTD
     for p in m.PERIODS:
         for lz in m.LOAD_ZONES:
-            m.BuildLocalTD[lz, p].setub(2 * m.zone_expected_coincident_peak_demand[lz, p])
+            m.BuildLocalTD[lz, p].setub(
+                2 * m.zone_expected_coincident_peak_demand[lz, p]
+            )
 
     # Estimate an upper bound of system peak demand for limiting generation unit
     # & transmission line builds
     system_wide_peak = {}
     for p in m.PERIODS:
         system_wide_peak[p] = sum(
-            m.zone_expected_coincident_peak_demand[lz, p] for lz in m.LOAD_ZONES)
+            m.zone_expected_coincident_peak_demand[lz, p] for lz in m.LOAD_ZONES
+        )
 
     # BuildGen
     for g, bld_yr in m.NEW_GEN_BLD_YRS:
@@ -27,6 +31,7 @@ def pysp_boundsetter_callback(self, scenario_tree, scenario):
     for tx, bld_yr in m.TRANS_BLD_YRS:
         m.BuildTx[tx, bld_yr].setub(5 * system_wide_peak[bld_yr])
 
+
 # For some reason runph looks for pysp_boundsetter_callback when run in
 # single-thread mode and ph_boundsetter_callback when called from mpirun with
 # remote execution via pyro. so we map both names to the same function.

examples/3zone_toy_stochastic_PySP/rhosetter.py

@@ -18,17 +18,22 @@
 from switch_model.utilities import iteritems
 
 try:
-    from pyomo.repn import generate_standard_repn    # Pyomo >=5.6
+    from pyomo.repn import generate_standard_repn  # Pyomo >=5.6
+
     newPyomo = True
 except ImportError:
-    from pyomo.repn import generate_canonical_repn   # Pyomo <=5.6
+    from pyomo.repn import generate_canonical_repn  # Pyomo <=5.6
+
     newPyomo = False
 
+
 def ph_rhosetter_callback(ph, scenario_tree, scenario):
-    # Derive coefficients from active objective    
-    cost_expr = next(scenario._instance.component_data_objects(
-        Objective, active=True, descend_into=True
-    ))
+    # Derive coefficients from active objective
+    cost_expr = next(
+        scenario._instance.component_data_objects(
+            Objective, active=True, descend_into=True
+        )
+    )
     set_rho_values(ph, scenario_tree, scenario, cost_expr)
 
 
@@ -57,8 +62,7 @@ def set_rho_values(ph, scenario_tree, scenario, cost_expr):
 
     cost_coefficients = {}
     var_names = {}
-    for (variable, coef) in \
-            zip(standard_repn.linear_vars, standard_repn.linear_coefs):
+    for (variable, coef) in zip(standard_repn.linear_vars, standard_repn.linear_coefs):
         variable_id = symbol_map.getSymbol(variable)
         cost_coefficients[variable_id] = coef
         var_names[variable_id] = variable.name
@@ -72,11 +76,13 @@ def set_rho_values(ph, scenario_tree, scenario, cost_expr):
                     tree_node,
                     scenario,
                     variable_id,
-                    cost_coefficients[variable_id] * rho_coefficient)
+                    cost_coefficients[variable_id] * rho_coefficient,
+                )
                 set_rho = True
                 break
         if not set_rho:
             print(
-                "Warning! Could not find tree node for variable {}; rho not set."
-                .format(var_names[variable_id])
+                "Warning! Could not find tree node for variable {}; rho not set.".format(
+                    var_names[variable_id]
+                )
             )

examples/3zone_toy_stochastic_PySP/rhosetter_FS_only.py

@@ -25,6 +25,7 @@
 # The rhosetter module should be in the same directory as this file.
 from rhosetter import set_rho_values
 
+
 def ph_rhosetter_callback(ph, scenario_tree, scenario):
     # This component name must match the expression used for first stage
     # costs defined in the ReferenceModel.