Bots And Games
战略决策
有了最新游戏状态 的辅助，您的机器人就可以进化为 自主代理。 这一转变标志着功能的升级，不仅支持对游戏状态的反应，还支持考虑上下文、能量和邻近度来制定决策的战略行动。
编写代码
返回到 bot.lua 文件并添加以下函数：
-- 确定两点之间的接近度。
function inRange(x1, y1, x2, y2, range)
return math.abs(x1 - x2) <= range and math.abs(y1 - y2) <= range
end
-- 根据距离和能量来战略性地决定下一步行动。
function decideNextAction()
local player = LatestGameState.Players[ao.id]
local targetInRange = false
for target, state in pairs(LatestGameState.Players) do
if target ~= ao.id and inRange(player.x, player.y, state.x, state.y, 1) then
targetInRange = true
break
end
end
if player.energy > 5 and targetInRange then
print("Player in range. Attacking.")
ao.send({Target = Game, Action = "PlayerAttack", Player = ao.id, AttackEnergy = tostring(player.energy)})
else
print("No player in range or insufficient energy. Moving randomly.")
local directionMap = {"Up", "Down", "Left", "Right", "UpRight", "UpLeft", "DownRight", "DownLeft"}
local randomIndex = math.random(#directionMap)
ao.send({Target = Game, Action = "PlayerMove", Player = ao.id, Direction = directionMap[randomIndex]})
end
end
decideNextAction 函数现在证明了我们的智能体基于对其环境的全面了解进行思考和行动的能力。 它会分析最新的游戏状态，如果您有足够的能量并且对手处于 inRange，则进行攻击，否则进行移动。
现在您所需要的只是一个处理程序来确保该函数自行运行。
Handlers.add(
"decideNextAction",
Handlers.utils.hasMatchingTag("Action", "UpdatedGameState"),
function ()
if LatestGameState.GameMode ~= "Playing" then
return
end
print("Deciding next action.")
decideNextAction()
end
)
该处理程序在收到已获取并更新最新游戏状态的消息时触发。 仅当游戏处于 Playing 模式时才会执行操作。
您可以在下面的代码中参考 bot.lua 的最新代码：
LatestGameState = LatestGameState or nil
function inRange(x1, y1, x2, y2, range)
return math.abs(x1 - x2) <= range and math.abs(y1 - y2) <= range
end
function decideNextAction()
local player = LatestGameState.Players[ao.id]
local targetInRange = false
for target, state in pairs(LatestGameState.Players) do
if target ~= ao.id and inRange(player.x, player.y, state.x, state.y, 1) then
targetInRange = true
break
end
end
if player.energy > 5 and targetInRange then
print("Player in range. Attacking.")
ao.send({Target = Game, Action = "PlayerAttack", Player = ao.id, AttackEnergy = tostring(player.energy)})
else
print("No player in range or insufficient energy. Moving randomly.")
local directionMap = {"Up", "Down", "Left", "Right", "UpRight", "UpLeft", "DownRight", "DownLeft"}
local randomIndex = math.random(#directionMap)
ao.send({Target = Game, Action = "PlayerMove", Player = ao.id, Direction = directionMap[randomIndex]})
end
end
Handlers.add(
"HandleAnnouncements",
Handlers.utils.hasMatchingTag("Action", "Announcement"),
function (msg)
ao.send({Target = Game, Action = "GetGameState"})
print(msg.Event .. ": " .. msg.Data)
end
)
Handlers.add(
"UpdateGameState",
Handlers.utils.hasMatchingTag("Action", "GameState"),
function (msg)
local json = require("json")
LatestGameState = json.decode(msg.Data)
ao.send({Target = ao.id, Action = "UpdatedGameState"})
end
)
Handlers.add(
"decideNextAction",
Handlers.utils.hasMatchingTag("Action", "UpdatedGameState"),
function ()
if LatestGameState.GameMode ~= "Playing" then
return
end
print("Deciding next action.")
decideNextAction()
end
)
加载和测试
再次，要测试最新的升级，请在 aos 玩家终端中加载文件，如下所示：
.load bot.lua
观察您的进程输出，以了解您的自主代理实时做出的决策，利用当前的游戏状态获得战略优势。 但是，如果在您决定下一步行动时另一个玩家攻击您并逃跑怎么办？ 在下一节中，您将学习在受到攻击后立即自动反击🤺