IGJSP 1.1.5__tar.gz → 1.1.6__tar.gz

{igjsp-1.1.5 → igjsp-1.1.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: IGJSP
-Version: 1.1.5
+Version: 1.1.6
 Summary: Instance generator for JSP
 Project-URL: Homepage, https://gps.blogs.upv.es/
 Author-email: GPS-UPV <gps@dsic.upv.es>

{igjsp-1.1.5 → igjsp-1.1.6}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "IGJSP"
-version = "1.1.5"
+version = "1.1.6"
 authors = [
   { name = "GPS-UPV", email = "gps@dsic.upv.es" },
 ]

{igjsp-1.1.5 → igjsp-1.1.6}/src/IGJSP/generador.py

@@ -3,6 +3,7 @@ import datetime
 import json
 import os
 import pickle
+import random
 import re
 from itertools import combinations
 
@@ -11,11 +12,11 @@ import numpy as np
 
 np.set_printoptions(linewidth=np.inf, threshold=np.inf, formatter={"int": lambda x: f"{x}"})
 
+from importlib.resources import as_file, files
 from pprint import pprint
 
 from scipy.stats import expon, norm, uniform
 
-from importlib.resources import files, as_file
 
 def _read_text_resource(rel_path: str, encoding: str = "utf-8") -> str:
     p = files("IGJSP").joinpath(rel_path)
@@ -150,17 +151,17 @@ class JSP:
         np.random.seed(seed)
         self.rddd = rddd
         self.speed = speed
-        #Elimino por que no hace caso del seed. Se tiene que hacer antes de llamar a esta función
-        # if not tpm or len(tpm) != self.numMchs:
-        #     if distribution == "uniform":
-        #         tpm = np.random.uniform(10, 100, self.numMchs)
-        #     elif distribution == "normal":
-        #         tpm = [max(10, data) for data in np.random.normal(50, 20, self.numMchs)]
-        #     else:
-        #         tpm = expon(loc=10, scale=20).rvs(self.numMchs)
+        if not tpm or len(tpm) != self.numMchs:
+            if distribution == "uniform":
+                tpm = np.random.uniform(10, 100, self.numMchs)
+            elif distribution == "normal":
+                tpm = [max(10, data) for data in np.random.normal(50, 20, self.numMchs)]
+            else:
+                tpm = expon(loc=10, scale=20).rvs(self.numMchs)
+       
         energyPer, timePer = self._particionate_speed_space(speed)
-        self._generate_standar_operation_cost(distribution,tpm)
-        
+        self._generate_standar_operation_cost(distribution)
+
         self.ProcessingTime = np.zeros((self.numJobs, self.numMchs, self.speed), dtype=int)
         self.EnergyConsumption = np.zeros((self.numJobs, self.numMchs, self.speed), dtype=int)
         self.Orden = np.zeros((self.numJobs, self.numMchs), dtype=int)
@@ -183,19 +184,17 @@ class JSP:
 
     def _particionate_speed_space(self, speed):
         energyPer = np.linspace(0.5, 3, speed) if speed > 1 else [1]
-        timePer = [t(c) for c in energyPer]
+        # timePer = [t(c) for c in energyPer]
+        timePer = sorted([random.uniform(0, 100)/100 for _ in range(3)])
         return energyPer, timePer
 
-    def _generate_standar_operation_cost(self, distribution,tpm=[]):
-        if np.array(tpm).shape != (self.numJobs, self.numMchs):
-            if distribution == "uniform":
-                self.operationCost = np.random.uniform(10, 100, (self.numJobs, self.numMchs))
-            elif distribution == "normal":
-                self.operationCost = np.array([max(10, x) for x in np.random.normal(50, 20, (self.numJobs, self.numMchs)).reshape(-1)]).reshape(self.numJobs, self.numMchs)
-            elif distribution == "exponential":
-                self.operationCost = np.random.exponential(10, (self.numJobs, self.numMchs))
-        else:
-            self.operationCost = tpm
+    def _generate_standar_operation_cost(self, distribution):
+        if distribution == "uniform":
+            self.operationCost = np.random.uniform(10, 100, (self.numJobs, self.numMchs))
+        elif distribution == "normal":
+            self.operationCost = np.array([max(10, x) for x in np.random.normal(50, 20, (self.numJobs, self.numMchs)).reshape(-1)]).reshape(self.numJobs, self.numMchs)
+        elif distribution == "exponential":
+            self.operationCost = np.random.exponential(10, (self.numJobs, self.numMchs))
 
     def _jobToMachine(self, release_date_tasks, timePer, distribution):
         for job in range(self.numJobs):
@@ -213,16 +212,18 @@ class JSP:
                     self.ReleaseDueDate[job, machine, 1] = releaseDateTask
                 else:
                     releaseDateTask += np.median(self.ProcessingTime[job, machine, :])
+            
             if self.rddd == 1:
                 self.ReleaseDueDate[job] = [initial, int(self._release_due(releaseDateTask, distribution))]
-    
 
     def _genProcEnergy(self, job, machine, timePer):        
-        ans = []  
-        for tper in timePer:
-            time = max(1, self.operationCost[job, machine] * tper)
-            ans.append((time, max(1, f(time))))
-        return ans
+        # ans = []  
+        # for tper in timePer:
+            # time = max(1, self.operationCost[job, machine] * tper)
+            # ans.append((time, max(1, f(time))))
+            # ans.append((time, max(1, f(time))))
+        timePer = sorted([random.uniform(0, 100)/100 for _ in range(3)])
+        return [(round(i*100),round((1-i)*100)) for i in timePer]
 
     def _release_due(self, duration, distribution):
         if distribution == "uniform":
@@ -666,19 +667,17 @@ class JSP:
             }
 
             if t == 1:
-                replace_data["releaseDate"] = str(data.ReleaseDueDate[:, 0].flatten()).replace(" ",", ")
-                # replace_data["releaseDate"] = ", ".join(list(map(str,data.ReleaseDueDate[:, 0])))
-                replace_data["dueDate"] = str(data.ReleaseDueDate[:, 1].flatten()).replace(" ",", ")
-                
+                replace_data["releaseDate"] = str([int(data.ReleaseDueDate[job, 0]) for job in range(data.numJobs)]).replace(", ", " ")
+                replace_data["dueDate"] = str([int(data.ReleaseDueDate[job, 1]) for job in range(data.numJobs)]).replace(", ", " ")
             elif t == 2:
-                replace_data["releaseDate"] = str(data.ReleaseDueDate[:,:, 0].flatten()).replace(" ",", ")
-                replace_data["dueDate"] = str(data.ReleaseDueDate[:, :, 1].flatten()).replace(" ",", ")
+                replace_data["releaseDate"] = str(data.ReleaseDueDate[:, :, 0].flatten()).replace(", ", " ")
+                replace_data["dueDate"] = str(data.ReleaseDueDate[:, :, 1].flatten()).replace(", ", " ")
 
             for job in range(data.numJobs):
                 for i, prioridad in enumerate(range(data.numMchs)):
                     precedence[job, data.Orden[job, prioridad]] = i
 
-            replace_data["precedence"] = str(precedence.flatten()).replace(" ", ", ")
+            replace_data["precedence"] = str(precedence.flatten()).replace(" ", ",")
             filedata = _read_text_resource(f"Minizinc/Types/RD/JSP/type{t}.dzn")
             # with open(f"./Minizinc/Types/RD/JSP/type{t}.dzn", "r", encoding="utf-8") as file:
                 # filedata = file.read()
@@ -1131,27 +1130,11 @@ class Generator:
                 jsp_instance = JSP(jobs = jobs, machines = machines)
             case "FJSP":
                 jsp_instance = FJSP(jobs = jobs, machines = machines)
-                
-        tpm_aux=[]
-        orden_aux=[]
-        for index in range(1, size + 1):
-            if len(tpm) != machines:
-                if distribution == "uniform":
-                    aux = np.random.uniform(10, 100, (jobs, machines))
-                elif distribution == "normal":
-                    aux = np.array([max(10, x) for x in np.random.normal(50, 20, (jobs, machines)).reshape(-1)]).reshape(jobs, machines)
-                elif distribution == "exponential":
-                    aux = np.random.exponential(10, (jobs, machines))
-            tpm_aux.append(aux)
-        
-            orden_aux.append([np.random.choice(range(machines), machines, replace=False) for job in range(jobs)])
-        
-        orden_aux = np.array(orden_aux)
-        
+        instances = []
         for index in range(1, size + 1):
+            
+            jsp_instance.fill_random_values(speed = speed, rddd = ReleaseDateDueDate, distribution = distribution, seed = seed,tpm = tpm)
 
-            jsp_instance.fill_random_values(speed = speed, rddd = ReleaseDateDueDate, distribution = distribution, seed = seed,tpm = tpm_aux[index-1])
-            jsp_instance.Orden = orden_aux[index-1]
             # Determinar el nombre de salida basado en `outputName` y los parámetros actuales
             problem_path = self.savepath.format(size = size, jobs =jobs, machines = machines, release_due_date = ReleaseDateDueDate,  speed_scaling = speed, distribution = distribution, seed=seed)
             
@@ -1183,5 +1166,5 @@ class Generator:
                     os.remove(pkl_path)
                 if self.taillard:
                     jsp_instance.saveTaillardStandardFile(f"{problem_path}/TAILLARD/" + jm_path.split("/")[0] + f"_{j}x{m}_{i}.txt")
-
-        return jsp_instance
+            instances.append(jsp_instance)
+        return instances if len(instances) >1 else instances[0]