import numpy as np
import pandas as pd

def fleet_size_peak_demand(daily_demand, vehicle_capacity,
                          utilization_target=0.85,
                          peak_percentile=95):
    """
    Calculate fleet size based on peak demand

    Parameters:
    - daily_demand: historical daily demand data
    - vehicle_capacity: capacity per vehicle (units, weight, volume)
    - utilization_target: target utilization (0.0-1.0)
    - peak_percentile: percentile for peak planning (e.g., 95)
    """

    # Calculate peak demand at specified percentile
    peak_demand = np.percentile(daily_demand, peak_percentile)

    # Calculate required fleet size
    fleet_size = np.ceil(peak_demand / (vehicle_capacity * utilization_target))

    # Calculate statistics
    avg_demand = np.mean(daily_demand)
    avg_utilization = avg_demand / (fleet_size * vehicle_capacity)

    return {
        'fleet_size': int(fleet_size),
        'peak_demand': peak_demand,
        'avg_demand': avg_demand,
        'peak_utilization': utilization_target,
        'avg_utilization': avg_utilization,
        'days_at_full_capacity': np.sum(daily_demand >= fleet_size * vehicle_capacity)
    }

# Example usage
daily_deliveries = np.random.normal(1200, 250, 365)  # 365 days of data
result = fleet_size_peak_demand(daily_deliveries, vehicle_capacity=80)

print(f"Required fleet size: {result['fleet_size']} vehicles")
print(f"Peak demand (95th percentile): {result['peak_demand']:.0f} deliveries")
print(f"Average utilization: {result['avg_utilization']:.1%}")

from scipy.stats import poisson
import math

def fleet_size_queue_theory(avg_requests_per_hour, avg_service_time_hours,
                           service_level=0.95):
    """
    Calculate fleet size using queue theory (M/M/c model)

    Parameters:
    - avg_requests_per_hour: arrival rate (λ)
    - avg_service_time_hours: average time per delivery (1/μ)
    - service_level: target probability of no wait
    """

    # Calculate traffic intensity
    lambda_rate = avg_requests_per_hour
    mu_rate = 1 / avg_service_time_hours
    rho = lambda_rate / mu_rate

    # Minimum servers (vehicles) needed
    min_servers = math.ceil(rho)

    # Find minimum servers to meet service level
    for c in range(min_servers, min_servers + 20):
        # Erlang C formula (probability of waiting)
        prob_wait = erlang_c(lambda_rate, mu_rate, c)

        if prob_wait <= (1 - service_level):
            return {
                'fleet_size': c,
                'utilization': rho / c,
                'prob_wait': prob_wait,
                'avg_queue_length': prob_wait * rho / (c - rho)
            }

    return {'fleet_size': min_servers + 20, 'message': 'Service level not achievable'}

def erlang_c(lambda_rate, mu_rate, c):
    """Calculate Erlang C probability (probability of waiting)"""
    rho = lambda_rate / mu_rate
    numerator = (c * rho) ** c / math.factorial(c)
    denominator = sum((c * rho) ** k / math.factorial(k) for k in range(c))
    denominator += (c * rho) ** c / (math.factorial(c) * (1 - rho / c))

    erlang_b = numerator / denominator
    erlang_c = erlang_b / (1 + erlang_b * (1 - rho / c))

    return erlang_c

import numpy as np
import simpy

class FleetSimulation:
    """
    Discrete-event simulation for fleet sizing

    Simulates daily operations to determine optimal fleet size
    """

    def __init__(self, env, num_vehicles):
        self.env = env
        self.vehicles = simpy.Resource(env, capacity=num_vehicles)
        self.completed_deliveries = 0
        self.total_wait_time = 0
        self.waiting_customers = 0

    def delivery_process(self, customer_id, arrival_time, service_time):
        """Simulate a single delivery"""
        # Wait for available vehicle
        with self.vehicles.request() as request:
            wait_start = self.env.now
            yield request

            wait_time = self.env.now - wait_start
            self.total_wait_time += wait_time

            if wait_time > 0:
                self.waiting_customers += 1

            # Perform delivery
            yield self.env.timeout(service_time)
            self.completed_deliveries += 1

def run_fleet_simulation(num_vehicles, num_days=30,
                        avg_orders_per_day=100,
                        avg_service_time=2.0):
    """
    Run fleet simulation

    Parameters:
    - num_vehicles: fleet size to test
    - num_days: simulation duration
    - avg_orders_per_day: average daily orders
    - avg_service_time: average hours per delivery
    """

    env = simpy.Environment()
    fleet = FleetSimulation(env, num_vehicles)

    # Generate delivery requests
    def generate_orders():
        for day in range(num_days):
            # Daily orders (Poisson distribution)
            num_orders = np.random.poisson(avg_orders_per_day)

            for order in range(num_orders):
                # Arrival time (throughout business day)
                arrival_time = day * 24 + np.random.uniform(8, 18)

                # Service time (lognormal distribution)
                service_time = np.random.lognormal(
                    mean=np.log(avg_service_time),
                    sigma=0.5
                )

                env.process(fleet.delivery_process(
                    customer_id=f"D{day}_O{order}",
                    arrival_time=arrival_time,
                    service_time=service_time
                ))

                yield env.timeout(0)

    env.process(generate_orders())
    env.run(until=num_days * 24)

    # Calculate metrics
    service_level = 1 - (fleet.waiting_customers / fleet.completed_deliveries)
    avg_wait = fleet.total_wait_time / fleet.completed_deliveries if fleet.completed_deliveries > 0 else 0
    utilization = fleet.completed_deliveries * avg_service_time / (num_vehicles * num_days * 24)

    return {
        'fleet_size': num_vehicles,
        'completed_deliveries': fleet.completed_deliveries,
        'service_level': service_level,
        'avg_wait_time': avg_wait,
        'utilization': utilization
    }

# Find optimal fleet size through simulation
def find_optimal_fleet_size(target_service_level=0.90):
    """Test different fleet sizes to find optimal"""

    results = []

    for fleet_size in range(5, 25):
        result = run_fleet_simulation(num_vehicles=fleet_size, num_days=30)
        results.append(result)

        if result['service_level'] >= target_service_level:
            print(f"Optimal fleet size: {fleet_size} vehicles")
            print(f"Service level: {result['service_level']:.2%}")
            print(f"Utilization: {result['utilization']:.2%}")
            return result

    return results

class VehicleTCO:
    """
    Calculate Total Cost of Ownership for vehicles

    Compares buy vs. lease scenarios
    """

    def __init__(self, vehicle_type):
        self.vehicle_type = vehicle_type

    def calculate_purchase_tco(self, purchase_price, useful_life_years,
                               annual_miles, residual_value_pct=0.20):
        """
        Calculate TCO for purchased vehicle

        Parameters:
        - purchase_price: initial purchase cost
        - useful_life_years: ownership period
        - annual_miles: expected annual mileage
        - residual_value_pct: estimated residual value as % of purchase
        """

        # Annual operating costs
        fuel_cost_per_mile = 0.35  # varies by vehicle type
        maintenance_per_mile = 0.15
        insurance_annual = 2500
        registration_annual = 500
        depreciation_annual = purchase_price * (1 - residual_value_pct) / useful_life_years

        # Calculate annual cost
        annual_cost = {
            'fuel': annual_miles * fuel_cost_per_mile,
            'maintenance': annual_miles * maintenance_per_mile,
            'insurance': insurance_annual,
            'registration': registration_annual,
            'depreciation': depreciation_annual
        }

        total_annual = sum(annual_cost.values())

        # Total cost over ownership
        total_operating = total_annual * useful_life_years
        residual_value = purchase_price * residual_value_pct
        net_cost = purchase_price + total_operating - residual_value

        # Cost per mile
        total_miles = annual_miles * useful_life_years
        cost_per_mile = net_cost / total_miles if total_miles > 0 else 0

        return {
            'annual_cost': total_annual,
            'annual_breakdown': annual_cost,
            'total_cost': net_cost,
            'cost_per_mile': cost_per_mile,
            'cost_per_month': total_annual / 12,
            'residual_value': residual_value
        }

    def calculate_lease_tco(self, monthly_lease_payment, lease_term_years,
                           annual_miles, excess_mile_charge=0.25):
        """
        Calculate TCO for leased vehicle

        Parameters:
        - monthly_lease_payment: monthly lease cost
        - lease_term_years: lease duration
        - annual_miles: expected annual mileage
        - excess_mile_charge: cost per mile over allowance
        """

        lease_allowance_miles = 15000  # typical annual allowance

        # Calculate excess miles
        excess_miles_per_year = max(0, annual_miles - lease_allowance_miles)
        excess_miles_total = excess_miles_per_year * lease_term_years
        excess_miles_cost = excess_miles_total * excess_mile_charge

        # Annual operating costs (lessee responsible)
        fuel_cost_per_mile = 0.35
        insurance_annual = 2500
        registration_annual = 500

        annual_operating = (
            annual_miles * fuel_cost_per_mile +
            insurance_annual +
            registration_annual
        )

        annual_lease_cost = monthly_lease_payment * 12

        # Total cost
        total_lease_payments = annual_lease_cost * lease_term_years
        total_operating = annual_operating * lease_term_years
        total_cost = total_lease_payments + total_operating + excess_miles_cost

        # Cost per mile
        total_miles = annual_miles * lease_term_years
        cost_per_mile = total_cost / total_miles if total_miles > 0 else 0

        return {
            'annual_cost': annual_lease_cost + annual_operating,
            'total_lease_payments': total_lease_payments,
            'excess_miles_cost': excess_miles_cost,
            'total_cost': total_cost,
            'cost_per_mile': cost_per_mile,
            'cost_per_month': (annual_lease_cost + annual_operating) / 12
        }

    def compare_buy_vs_lease(self, purchase_price, lease_monthly,
                            years, annual_miles):
        """
        Compare buying vs. leasing

        Returns comparison with recommendation
        """

        buy_tco = self.calculate_purchase_tco(
            purchase_price=purchase_price,
            useful_life_years=years,
            annual_miles=annual_miles
        )

        lease_tco = self.calculate_lease_tco(
            monthly_lease_payment=lease_monthly,
            lease_term_years=years,
            annual_miles=annual_miles
        )

        savings = lease_tco['total_cost'] - buy_tco['total_cost']
        recommendation = 'Buy' if savings > 0 else 'Lease'

        return {
            'buy': buy_tco,
            'lease': lease_tco,
            'savings_with_buy': savings,
            'recommendation': recommendation,
            'buy_cost_per_mile': buy_tco['cost_per_mile'],
            'lease_cost_per_mile': lease_tco['cost_per_mile']
        }

# Example usage
tco = VehicleTCO('Delivery Van')
comparison = tco.compare_buy_vs_lease(
    purchase_price=45000,
    lease_monthly=650,
    years=5,
    annual_miles=25000
)

print(f"Recommendation: {comparison['recommendation']}")
print(f"Buy TCO: ${comparison['buy']['total_cost']:,.0f}")
print(f"Lease TCO: ${comparison['lease']['total_cost']:,.0f}")
print(f"Savings with buy: ${comparison['savings_with_buy']:,.0f}")

import numpy as np
from scipy.optimize import minimize_scalar

def optimal_replacement_age(purchase_price, annual_depreciation,
                           annual_maintenance, discount_rate=0.08):
    """
    Calculate optimal vehicle replacement age

    Minimizes equivalent annual cost (EAC)

    Parameters:
    - purchase_price: initial cost
    - annual_depreciation: depreciation schedule (list)
    - annual_maintenance: maintenance cost schedule (list)
    - discount_rate: discount rate for NPV
    """

    def equivalent_annual_cost(age):
        """Calculate EAC for given replacement age"""

        if age < 1 or age > len(annual_maintenance):
            return float('inf')

        age = int(age)

        # Calculate NPV of costs
        npv_costs = purchase_price

        for year in range(1, age + 1):
            discount_factor = (1 + discount_rate) ** -year
            npv_costs += annual_maintenance[year - 1] * discount_factor

        # Calculate residual value
        residual_value = purchase_price
        for year in range(age):
            residual_value -= annual_depreciation[year]

        npv_costs -= residual_value * (1 + discount_rate) ** -age

        # Convert to EAC
        annuity_factor = (
            (discount_rate * (1 + discount_rate) ** age) /
            ((1 + discount_rate) ** age - 1)
        )
        eac = npv_costs * annuity_factor

        return eac

    # Find optimal age
    result = minimize_scalar(
        equivalent_annual_cost,
        bounds=(1, len(annual_maintenance)),
        method='bounded'
    )

    optimal_age = int(result.x)
    min_eac = result.fun

    return {
        'optimal_replacement_age': optimal_age,
        'equivalent_annual_cost': min_eac
    }

# Example: Analyze replacement for delivery truck
purchase_price = 50000
annual_depreciation = [10000, 8000, 6000, 5000, 4000, 3000, 2000]
annual_maintenance = [2000, 2500, 3000, 4000, 5500, 7000, 9000]

result = optimal_replacement_age(
    purchase_price,
    annual_depreciation,
    annual_maintenance
)

print(f"Optimal replacement age: {result['optimal_replacement_age']} years")
print(f"Equivalent annual cost: ${result['equivalent_annual_cost']:,.0f}")

class FleetReplacementPlanner:
    """
    Plan multi-year fleet replacement strategy

    Accounts for budget constraints and vehicle ages
    """

    def __init__(self, fleet_data, annual_budget):
        """
        Parameters:
        - fleet_data: DataFrame with columns ['vehicle_id', 'age', 'type',
                      'replacement_cost', 'maintenance_cost']
        - annual_budget: maximum annual replacement budget
        """
        self.fleet = fleet_data.copy()
        self.annual_budget = annual_budget

    def prioritize_replacements(self, current_year):
        """
        Prioritize vehicles for replacement

        Scoring based on age, maintenance cost, and utilization
        """

        # Calculate replacement scores
        self.fleet['age_score'] = self.fleet['age'] / 10  # normalize
        self.fleet['maintenance_score'] = (
            self.fleet['maintenance_cost'] /
            self.fleet['maintenance_cost'].median()
        )

        # Combined score (higher = more urgent)
        self.fleet['replacement_score'] = (
            0.6 * self.fleet['age_score'] +
            0.4 * self.fleet['maintenance_score']
        )

        # Sort by priority
        self.fleet = self.fleet.sort_values(
            'replacement_score',
            ascending=False
        )

        return self.fleet[['vehicle_id', 'age', 'replacement_score']]

    def create_replacement_plan(self, planning_horizon=5):
        """
        Create multi-year replacement plan

        Returns year-by-year replacement schedule
        """

        replacement_plan = {year: [] for year in range(planning_horizon)}
        remaining_fleet = self.fleet.copy()

        for year in range(planning_horizon):
            # Age all vehicles by one year
            remaining_fleet['age'] += 1

            # Re-prioritize
            remaining_fleet['age_score'] = remaining_fleet['age'] / 10
            remaining_fleet['replacement_score'] = (
                0.6 * remaining_fleet['age_score'] +
                0.4 * remaining_fleet['maintenance_score']
            )
            remaining_fleet = remaining_fleet.sort_values(
                'replacement_score',
                ascending=False
            )

            # Select vehicles to replace within budget
            budget_remaining = self.annual_budget
            year_replacements = []

            for idx, vehicle in remaining_fleet.iterrows():
                if budget_remaining >= vehicle['replacement_cost']:
                    year_replacements.append({
                        'vehicle_id': vehicle['vehicle_id'],
                        'age_at_replacement': vehicle['age'],
                        'cost': vehicle['replacement_cost']
                    })
                    budget_remaining -= vehicle['replacement_cost']

                    # Remove from fleet
                    remaining_fleet = remaining_fleet.drop(idx)

            replacement_plan[year] = year_replacements

        return replacement_plan

    def summarize_plan(self, plan):
        """Create summary of replacement plan"""

        summary = []
        for year, replacements in plan.items():
            total_cost = sum(r['cost'] for r in replacements)
            num_vehicles = len(replacements)

            summary.append({
                'year': year,
                'num_replacements': num_vehicles,
                'total_cost': total_cost,
                'budget_utilization': total_cost / self.annual_budget
            })

        return pd.DataFrame(summary)

# Example usage
import pandas as pd

fleet_data = pd.DataFrame({
    'vehicle_id': [f'V{i:03d}' for i in range(50)],
    'age': np.random.randint(1, 12, 50),
    'type': np.random.choice(['Van', 'Truck', 'Tractor'], 50),
    'replacement_cost': np.random.randint(35000, 75000, 50),
    'maintenance_cost': np.random.randint(2000, 8000, 50)
})

planner = FleetReplacementPlanner(fleet_data, annual_budget=500000)
plan = planner.create_replacement_plan(planning_horizon=5)
summary = planner.summarize_plan(plan)

print("5-Year Replacement Plan:")
print(summary)

class FleetUtilizationAnalyzer:
    """
    Analyze and improve fleet utilization

    Key metrics: miles driven, hours used, revenue per vehicle
    """

    def __init__(self, vehicle_data, operational_data):
        """
        Parameters:
        - vehicle_data: DataFrame with vehicle details
        - operational_data: DataFrame with daily usage logs
        """
        self.vehicles = vehicle_data
        self.operations = operational_data

    def calculate_utilization_metrics(self):
        """Calculate comprehensive utilization metrics"""

        metrics = []

        for vehicle_id in self.vehicles['vehicle_id']:
            vehicle_ops = self.operations[
                self.operations['vehicle_id'] == vehicle_id
            ]

            if len(vehicle_ops) == 0:
                continue

            # Time utilization
            total_days = (vehicle_ops['date'].max() -
                         vehicle_ops['date'].min()).days + 1
            active_days = vehicle_ops['date'].nunique()
            time_utilization = active_days / total_days if total_days > 0 else 0

            # Mileage
            total_miles = vehicle_ops['miles'].sum()
            avg_daily_miles = total_miles / active_days if active_days > 0 else 0

            # Revenue
            total_revenue = vehicle_ops['revenue'].sum()
            revenue_per_mile = total_revenue / total_miles if total_miles > 0 else 0

            # Efficiency
            empty_miles = vehicle_ops['empty_miles'].sum()
            loaded_miles = total_miles - empty_miles
            loaded_percentage = loaded_miles / total_miles if total_miles > 0 else 0

            metrics.append({
                'vehicle_id': vehicle_id,
                'active_days': active_days,
                'time_utilization': time_utilization,
                'total_miles': total_miles,
                'avg_daily_miles': avg_daily_miles,
                'total_revenue': total_revenue,
                'revenue_per_mile': revenue_per_mile,
                'loaded_percentage': loaded_percentage
            })

        return pd.DataFrame(metrics)

    def identify_underutilized_vehicles(self, utilization_threshold=0.60):
        """
        Identify vehicles below utilization threshold

        Returns candidates for removal from fleet
        """

        metrics = self.calculate_utilization_metrics()

        underutilized = metrics[
            metrics['time_utilization'] < utilization_threshold
        ].sort_values('time_utilization')

        return underutilized

    def recommend_fleet_adjustments(self):
        """
        Recommend fleet size adjustments

        Based on utilization analysis
        """

        metrics = self.calculate_utilization_metrics()

        avg_utilization = metrics['time_utilization'].mean()
        underutilized = len(metrics[metrics['time_utilization'] < 0.60])

        recommendations = []

        if avg_utilization < 0.65:
            potential_reduction = int(len(metrics) * (0.65 - avg_utilization))
            recommendations.append({
                'action': 'Reduce fleet size',
                'vehicles': potential_reduction,
                'reason': f'Average utilization only {avg_utilization:.1%}'
            })

        if underutilized > 0:
            recommendations.append({
                'action': 'Remove or reassign',
                'vehicles': underutilized,
                'reason': f'{underutilized} vehicles below 60% utilization'
            })

        # Check for overutilization
        high_mileage = metrics[metrics['avg_daily_miles'] > 250]
        if len(high_mileage) > len(metrics) * 0.3:
            recommendations.append({
                'action': 'Add vehicles',
                'vehicles': 2,
                'reason': 'Many vehicles exceeding 250 miles/day'
            })

        return recommendations

class MaintenanceScheduler:
    """
    Schedule and track fleet maintenance

    Prevents breakdowns, optimizes maintenance timing
    """

    def __init__(self, fleet_data):
        self.fleet = fleet_data.copy()

    def calculate_maintenance_schedule(self, current_date):
        """
        Calculate upcoming maintenance needs

        Returns schedule for next 90 days
        """

        schedule = []

        for idx, vehicle in self.fleet.iterrows():
            # Check mileage-based maintenance
            miles_since_service = (
                vehicle['current_mileage'] -
                vehicle['last_service_mileage']
            )

            miles_to_service = vehicle['service_interval'] - miles_since_service
            avg_daily_miles = vehicle['avg_daily_miles']

            if avg_daily_miles > 0:
                days_to_service = miles_to_service / avg_daily_miles
            else:
                days_to_service = 999

            # Check time-based maintenance
            days_since_service = (
                current_date - vehicle['last_service_date']
            ).days

            days_to_time_service = 180 - days_since_service  # 6-month interval

            # Use whichever comes first
            days_to_service = min(days_to_service, days_to_time_service)

            if days_to_service <= 90:  # Next 90 days
                schedule.append({
                    'vehicle_id': vehicle['vehicle_id'],
                    'days_to_service': max(0, days_to_service),
                    'due_date': current_date + pd.Timedelta(days=max(0, days_to_service)),
                    'service_type': 'Preventive Maintenance',
                    'estimated_cost': vehicle['avg_service_cost'],
                    'priority': 'High' if days_to_service <= 7 else 'Medium'
                })

        return pd.DataFrame(schedule).sort_values('days_to_service')

    def optimize_maintenance_timing(self, schedule, routes_data):
        """
        Optimize maintenance timing to minimize service disruption

        Coordinate with route schedules
        """

        optimized = []

        for idx, maintenance in schedule.iterrows():
            vehicle_id = maintenance['vehicle_id']

            # Find days when vehicle has lightest workload
            vehicle_routes = routes_data[
                routes_data['vehicle_id'] == vehicle_id
            ]

            # Calculate workload by day
            daily_workload = vehicle_routes.groupby('date').agg({
                'stops': 'sum',
                'miles': 'sum'
            })

            # Find lightest day within maintenance window
            maintenance_window_start = maintenance['due_date'] - pd.Timedelta(days=7)
            maintenance_window_end = maintenance['due_date'] + pd.Timedelta(days=3)

            window_workload = daily_workload[
                (daily_workload.index >= maintenance_window_start) &
                (daily_workload.index <= maintenance_window_end)
            ]

            if len(window_workload) > 0:
                optimal_date = window_workload['stops'].idxmin()
            else:
                optimal_date = maintenance['due_date']

            maintenance['optimal_date'] = optimal_date
            optimized.append(maintenance)

        return pd.DataFrame(optimized)